Transportation route planning method and system for large-scale transportation vehicles based on AI algorithm

Abstract

The present application relates to a method, system, computer equipment, storage medium and computer program product for planning the transportation route of large-scale transport vehicles based on AI algorithms. The method includes: building a database of route traffic information and transportation hub information in the region; obtaining the starting point, end point and vehicle information of large-scale transport vehicles; according to the starting point and end point, converting the routes and transportation hubs in the database into a graph structure in a graph neural network, generating a graph structure and graph structure information of the target area; the graph structure uses the interconnection hubs between roads as nodes, and the roads between the interconnection hubs as edges; the graph structure information includes node feature information and edge feature information; the starting point, end point, vehicle information, the graph structure of the target area and the graph structure information are input into a pre-trained prediction model, and the optimal transportation route of the large-scale transport vehicle is output through the prediction model, and the prediction model is a neural network model. This method can efficiently generate the optimal transportation route.

Description

Technical Field

[0001] The present application relates to the field of path planning technology and artificial intelligence technology, and in particular to a method, system, computer equipment, storage medium and computer program product for transport path planning of large-scale transport vehicles based on AI algorithms. Background Art

[0002] As the country vigorously encourages the rapid development of high-end manufacturing industries such as aerospace, wind power, and electromechanical, the demand for large-scale transportation is also increasing year by year. The main objects of large-scale transportation include wind turbine blades, shield machines, large excavators, large transformers, etc., which are characterized by "overweight, overlength, overwidth, overheight, high value, and cannot be disassembled."

[0003] However, due to height and width restrictions, some highway bridges cannot be used by heavy-duty transport vehicles, and the safety of heavy-duty transport vehicles passing through highway bridges due to their overweight loads needs to be verified and evaluated. On the one hand, the passability of the transportation route for heavy-duty transport is not only about the time cost of the route, but also involves highway bridge access restrictions and the safety of highway bridges; on the other hand, the highway bridge verification and evaluation process is a large-scale, repetitive and time-consuming process.

[0004] Therefore, there is an urgent need for a safe and efficient method for automatically and quickly generating transportation routes for large-scale transportation vehicles. Summary of the invention

[0005] Based on this, it is necessary to provide a safe and efficient AI-based large-scale transport vehicle transportation route planning method, system, computer equipment, computer-readable storage medium and computer program product to address the above-mentioned technical problems.

[0006] In a first aspect, the present application provides a method for transport path planning of large-scale transport vehicles based on an AI algorithm. The method comprises:

[0007] Build a database of route traffic information and transportation hub information in the region;

[0008] Obtain the starting point, end point and vehicle information of large-scale transport vehicles;

[0009] According to the starting point and the end point, the routes and transportation hubs in the database are converted into a graph structure in a graph neural network to generate a graph structure and graph structure information of the target area; the graph structure uses the interconnection hubs between the roads as nodes and the roads between the interconnection hubs as edges; the graph structure information includes node feature information and edge feature information; the node feature information includes the traffic restriction information of the node, and the edge feature information includes the traffic safety information of the road; the traffic safety information includes whether the road meets the traffic requirements of large-scale transport vehicles, or the safety factor and precautions for the passage of large-scale transport vehicles;

[0010] The starting point, the end point, the vehicle information, the graph structure of the target area and the graph structure information are input into a pre-trained prediction model, and the optimal transportation path of the large-scale transportation vehicle is output through the prediction model, and the prediction model is a neural network model.

[0011] In one embodiment, the traffic restriction information of the node includes: vehicle height restriction, vehicle length restriction, driving width restriction, and maximum load restriction.

[0012] In one of the embodiments, the traffic safety information also includes the traffic safety information of large-scale transport vehicles on the road; the training loss of the prediction model includes three basic constraints: nodes and paths that exceed the traffic limit are not within the path range; minimization of the planned path length, and maximization of the traffic safety of the traffic path.

[0013] In one embodiment, the method further comprises:

[0014] When the graph structure and graph structure information of the update area are obtained, the prediction model is trained according to the graph structure and graph structure information of the update area to update the prediction model.

[0015] In one embodiment, the prediction model uses a graph neural convolutional network; the convolutional layer of the graph neural convolutional network can consider the weight influence of nodes with edge connections when processing the node information depth, and the calculation formula is:

[0016] ;

[0017] in, and is the fully connected layer (MLP) of the neural network; x i (k-1) and x i (k) They are k -1st floor and next floor k Feature vector of layer data points; is the edge connection eigenvector, x j (k-1) is the corresponding feature vector of nearby data points.

[0018] In one embodiment, the graph neural convolution network adopts a graph convolution aggregation method, and its node information depth aggregation relationship is:

[0019] ;

[0020] in, H (l) and H (l+1) They are the node information representations before and after aggregation respectively; represents the activation function; Represents the adjacency matrix between nodes; yes The degree matrix of W (l) yes l Layer and l +1 Neural network training parameters between layers.

[0021] In a second aspect, the present application also provides a large-scale transport vehicle transportation path planning system based on an AI algorithm. The system includes:

[0022] Construction module, used to build a database of route traffic information and transportation hub information in the region;

[0023] Planning acquisition module, used to obtain the starting point, end point and vehicle information of large-scale transport vehicles;

[0024] A graph information acquisition module is used to convert the routes and transportation hubs in the database into a graph structure in a graph neural network according to the starting point and the end point, and generate a graph structure and graph structure information of the target area; the graph structure uses the interconnection hubs between the roads as nodes, and the roads between the interconnection hubs as edges; the graph structure information includes node feature information and edge feature information; the node feature information includes the traffic restriction information of the node, and the edge feature information includes the traffic safety information of the road; the traffic safety information includes whether the road meets the traffic requirements of large-scale transport vehicles, or the safety factor and precautions for the passage of large-scale transport vehicles;

[0025] The prediction module is used to input the starting point, the end point, the vehicle information, the graph structure of the target area and the graph structure information into a pre-trained prediction model, and output the optimal transportation path of the large-scale transport vehicle through the prediction model. The prediction model is a neural network model.

[0026] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-mentioned embodiments of the method for planning a transportation path for a large-scale transportation vehicle based on an AI algorithm.

[0027] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned methods for planning a transportation path for large-scale transportation vehicles based on an AI algorithm.

[0028] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for planning a transportation path for a large-scale transportation vehicle based on an AI algorithm are implemented.

[0029] The above-mentioned AI algorithm-based large-scale transport vehicle transportation route planning method, system, computer equipment, storage medium and computer program product express the vehicle traffic information data in the area in the form of a graph structure, and train the corresponding graph neural network model to plan the optimal transportation route for large-scale transport vehicles. The optimal transportation route can be efficiently generated without repeated manual comparisons. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is an application environment diagram of a method for planning a transportation path for a large-scale transportation vehicle based on an AI algorithm in one embodiment;

[0031] Figure 2 A schematic diagram of a flow chart of a method for planning a transportation route for a large-scale transportation vehicle based on an AI algorithm in one embodiment;

[0032] Figure 3 is a schematic diagram of a graph structure of an embodiment;

[0033] Figure 4 A schematic diagram illustrating a graph neural network computing framework according to an embodiment;

[0034] Figure 5 A schematic diagram of a graph neural network framework according to an embodiment;

[0035] Figure 6 A flowchart of a method for planning a transportation route for a large-scale transportation vehicle based on an AI algorithm according to an embodiment of the present invention;

[0036] Figure 7 A structural block diagram of a transport path planning system for large-scale transport vehicles based on an AI algorithm according to an embodiment;

[0037] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0039] The method for planning the transportation path of a large-scale transport vehicle based on an AI algorithm provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 is a department approval terminal for large-scale transport vehicles, and the server 104 can be implemented with an independent server or a server cluster composed of multiple servers.

[0040] In one embodiment, Figure 2 As shown in the figure, a transportation path planning method for large-scale transportation vehicles based on AI algorithm is provided. Figure 1 The server in the example is used to illustrate the following steps:

[0041] Step 200, building a database of route traffic information and transportation hub information in the region.

[0042] Step 202, obtaining the starting point, end point and vehicle information of the large-scale transport vehicle.

[0043] Oversized transport refers to the oversized transport of indivisible items. Oversized transport vehicles refer to cargo transport vehicles that have any of the following conditions:

[0044] The total height of the vehicle and cargo exceeds 4 meters from the ground;

[0045] The total width of the vehicle and cargo exceeds 2.55 meters;

[0046] The total length of the vehicle and cargo exceeds 18.1 meters;

[0047] The total mass of the vehicle and cargo exceeds 49 tons.

[0048] The vehicle information may include information about cargo loaded on the vehicle, vehicle weight, dimensions, and other information.

[0049] Step 204, according to the starting point and the end point, the routes and transportation hubs in the database are converted into a graph structure in the graph neural network, and the graph structure and graph structure information of the target area are generated; the graph structure uses the interchange hubs between roads as nodes, and the roads between the interchange hubs as edges; the graph structure information includes node feature information and edge feature information; the node feature information includes the traffic restriction information of the node, and the edge feature information includes the traffic safety information of the road; the traffic safety information includes whether the road meets the traffic requirements of large-scale transport vehicles, or the safety factor and precautions for the passage of large-scale transport vehicles.

[0050] Among them, the system has a map library, and the map library can be divided into administrative units, such as maps of multiple cities, or maps of multiple provinces. According to the starting point and the end point, the target area is matched from the map library. It should be understood that the target area includes the administrative unit where the starting point is located, the administrative unit where the end point is located, and the administrative unit from the starting point to the end point. For example, if the starting point is Changsha City, Hunan Province, and the end point is Wuhan City, Hubei Province, then the target area can include two provinces, Hunan Province and Hubei Province.

[0051] According to the starting point and the end point, the routes and transportation hubs in the database are converted into the graph structure in the graph neural network to generate the graph structure and graph structure information of the target area. The graph structure consists of two features: nodes and edges. Nodes and edges have corresponding attribute vectors. The vectors can be Boolean vectors (0 / 1, i.e., yes or no) or specific numerical values. In one embodiment, the graph structure is as follows: Figure 3 shown.

[0052] In this embodiment, the graph structure uses the interconnection hubs between roads as nodes, and the roads between the interconnection hubs as edges; the graph structure information includes node feature information and edge feature information. For example, the interconnection hubs of highways can be used as nodes, and the highways between the interconnection hubs can be used as edges to construct a map.

[0053] The graph structure information of the target area includes node feature information and edge feature information. Node feature information includes the traffic restriction information of large-scale transport vehicles at the node. For example, it includes vehicle height, vehicle length, driving width and maximum load limit. Edge feature information includes whether the road is a passage for large-scale transport vehicles and the traffic safety information of large-scale transport vehicles on the road. Among them, it can be preliminarily judged whether the road can be used as a passage for large-scale transport vehicles based on the bridge inspection level, minimum linear radius, mileage, bridge bearing capacity, etc., and a probabilistic traffic safety factor is given to the road based on the existing bridge verification structure, and precautions can also be included.

[0054] Step 206, input the starting point, the end point, the vehicle information, the graph structure of the target area and the graph structure information into a pre-trained prediction model, and output the optimal transportation path of the large-scale transportation vehicle through the prediction model, where the prediction model is a neural network model.

[0055] At present, the route planning problem for large-scale transportation is still at the stage of manual route comparison, and a systematic and rigorous scientific prediction system has not been formed. On the basis of setting the starting point and end point of large-scale transportation, according to the conventional route navigation, if there are restrictions on local routes and they cannot be passed, the local routes will be adjusted, and then the highway bridges of the entire route will be verified and evaluated. If some of the bridges fail to pass the verification, the route will be adjusted locally in the same way. The entire route planning process is repeated through adjustment, verification, update and optimization until a transportation route that can meet the requirements of large-scale transportation is obtained.

[0056] At present, the volume of large-scale transportation is increasing, but the route planning of large-scale transportation has not yet formed a scientific, efficient and systematic route generation method. In summary, the current large-scale transportation route output still relies on manual repeated comparisons, which is inefficient.

[0057] In this embodiment, a neural network model is used to plan the transportation path of large-scale transportation vehicles. In this embodiment, the neural network model can adopt a computational framework of a graph neural network. Figure 4 The computing framework of a graph neural network of an embodiment is composed of three parts: an input layer, an intermediate deep hidden layer, and an output layer. Among them, the input layer provides the feature vector of the global node, the edge connection between nodes, and the feature vector of the edge; the output layer can output the prediction vector of the node or the prediction vector of the edge, and the relationship established by the multiple hidden layers between the input and output is a function model built by multiple trainable variable parameters.

[0058] By pre-training the prediction model, the transportation routes of large-scale transport vehicles can be planned, and the optimal transportation route can be efficiently generated without repeated manual comparisons.

[0059] The above-mentioned transportation route planning method for large-scale transport vehicles based on AI algorithm expresses the vehicle traffic information data in the area in the form of a graph structure and trains the corresponding graph neural network model to plan the optimal transportation route for large-scale transport vehicles. The optimal transportation route can be efficiently generated without repeated manual comparisons.

[0060] In one embodiment, the training loss of the prediction model includes three basic constraints.

[0061] Among them, the first basic constraint condition is: nodes and paths that exceed the access limit are not within the path range.

[0062] Traffic restrictions can include two aspects: one is whether the traffic restrictions are exceeded; the other is whether the road section is a passage for large transport vehicles.

[0063] Specifically, the graph structure information includes node feature information, and the node feature information includes node access restriction information. The node access restriction information may include vehicle height restriction, vehicle length restriction, driving width restriction, and maximum load restriction. Specifically, whether a certain node in the graph node is within the path range can be determined based on the degree of matching between the vehicle information and the node access restriction information.

[0064] Specifically, the graph structure information includes edge feature information, and the edge feature information includes whether the road is a passage for large-scale transport vehicles. If the road is not a passage for large-scale transport vehicles, the road cannot be planned as a transport route.

[0065] Through the first basic constraint condition, nodes that exceed the access limit are excluded from the path range, and routes that are not for large-scale transport vehicles are excluded.

[0066] Among them, the second basic constraint condition is: minimizing the length of the planned path. That is, the second basic constraint condition constrains the planned path so that the transportation path output by the prediction model is the shortest.

[0067] The third basic constraint is: maximize the safety of the path. That is, the third basic constraint constrains from the aspect of safety, so that the transportation path output by the prediction model is the safest. Specifically, for a planned path, the safety factor of large-scale transport vehicles on the road is obtained from the side information of the target area, and the safety factor of each road on the path is multiplied or added to obtain the path with maximum safety.

[0068] In this embodiment, the prediction model is trained through the above three basic constraints, so that the prediction model can output a multi-objective optimal transportation path that meets the traffic restrictions, minimizes the planned path length, and maximizes the traffic safety of the traffic path.

[0069] In another embodiment, the prediction model can be made scalable through data collection and updating. Specifically, when the graph structure and graph structure information of the update area are obtained, the prediction model is trained according to the graph structure and graph structure information of the update area to update the prediction model.

[0070] For example, if the data area is currently limited to Hunan Province, when the hub and highway bridge data in Hubei Province are updated, route planning in the two lakes area can be realized after model training. The prediction model can be updated by training the prediction model with the updated map graph structure, so that the prediction model has growth and scalability.

[0071] In this embodiment, as the target traffic area and highway bridge information are updated, the prediction model can be updated and upgraded in a timely manner.

[0072] In another embodiment, the prediction model uses a graph neural convolutional network; the convolutional layer of the graph neural convolutional network can consider the weight influence of the nodes with which it has edge connections when processing the node information depth, and its calculation is shown in formula (1):

[0073] (1)

[0074] in, and is the fully connected layer (MLP) of the neural network; x i (k-1) and x i (k) They are k -1st floor and next floor k Feature vector of layer data points; is the edge connection eigenvector, x j (k-1) is the corresponding feature vector of nearby data points.

[0075] This method uses the graph convolution (GCN) aggregation method, and its node information depth aggregation relationship is shown in formula (2):

[0076] (2)

[0077] in, H (l) and H (l+1) They are the node information representations before and after aggregation respectively; represents the activation function; Represents the adjacency matrix between nodes; yes The degree matrix of W (l) yes l Layer and l +1 Neural network training parameters between layers.

[0078] A graph neural network framework of an embodiment is as follows Figure 5 As shown in the figure, the graph neural network framework is trained based on three basic constraints, including the perception layer (MLP), convolution layer (GCN), and aggregation layer. The loss function Loss is composed of three basic constraints: (1) nodes and paths that exceed the pass limit are not within the path range; (2) the length of the planned path is minimized; (3) the process safety factor of the pass path is maximized.

[0079] In one embodiment, a method for planning a transportation route for a large-scale transportation vehicle based on an AI algorithm is provided. Figure 6 As shown, it includes three stages:

[0080] The first stage is data preparation.

[0081] Specifically, in the first stage, we need to prepare the node information and edge information required to build the graph structure. Nodes are transportation hubs, and edges are highway bridges between hubs.

[0082] Node information includes vehicle height limit, vehicle length limit, driving width limit, and maximum load limit.

[0083] The edge information includes whether the road is a passage for large-scale transport vehicles and the safety information of large-scale transport vehicles on the road.

[0084] The second stage is model training.

[0085] Among them, model training is the core part of the whole process, which can be divided into two parts:

[0086] (1) Data characterization (extracting data into a graph structure)

[0087] The interchange hubs between highways are taken as nodes, and the characteristic information assigned by the nodes may include vehicle height, vehicle length, driving width, and maximum load limit. The nodes are used to divide the highway line into multiple side connections, and the side connections are assigned with traffic information (bridge inspection level, minimum linear radius, mileage, bridge bearing capacity, etc.) to preliminarily determine whether the section can be used as a route for large-scale transport vehicles, and a probabilistic traffic safety factor is assigned based on the existing bridge verification structure.

[0088] (2) Neural network framework

[0089] The function of the convolutional layer of the graph neural network is to consider the weight influence of the nodes with which it has edge connections when processing the node's information depth. Its calculation is shown in formula (1):

[0090] (1)

[0091] in, and is the fully connected layer (MLP) of the neural network; x i (k-1) and x i (k) They are k -1st floor and next floor k Feature vector of layer data points; is the edge connection eigenvector, x j (k-1) is the corresponding feature vector of nearby data points.

[0092] This method uses the graph convolution (GCN) aggregation method, and its node information depth aggregation relationship is shown in formula (2):

[0093] (2)

[0094] in, H (l) and H (l+1) They are the node information representations before and after aggregation respectively; represents the activation function; Represents the adjacency matrix between nodes; yes The degree matrix of W (l) yes l Layer and l +1 Neural network training parameters between layers.

[0095] The loss function Loss is composed of three basic constraints: (1) nodes that exceed the access limit are not within the path range; (2) the length of the planned path is minimized; and (3) the process safety factor of the access path is maximized.

[0096] The third stage is path prediction.

[0097] Specifically, by inputting the starting point, terminal and vehicle information, a transportation route is generated using a prediction model.

[0098] In this application, in view of the particularity of large-scale transportation route planning, an AI model is used to form a systematic and scientific route planning method. At the same time, the route planning model based on AI deep learning is growable and scalable. With the update of the target traffic area and highway bridge information, the model can be updated and upgraded in time. Among them, the AI ​​deep learning model has the ability to learn complex bridge verification features.

[0099] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0100] Based on the same inventive concept, the embodiment of the present application also provides a transport path planning system for large-scale transport vehicles for implementing the above-mentioned transport path planning method for large-scale transport vehicles based on AI algorithms. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more embodiments of the transport path planning system for large-scale transport vehicles based on AI algorithms provided below can be referred to the limitations of the transport path planning method for large-scale transport vehicles based on AI algorithms above, and will not be repeated here.

[0101] In one embodiment, Figure 7 As shown, a large-scale transport vehicle transportation path planning system based on AI algorithm is provided, including:

[0102] Construction module 700 is used to build a database of route traffic information and transportation hub information in the region.

[0103] The planning acquisition module 702 is used to obtain the starting point, end point and vehicle information of the large-scale transport vehicle.

[0104] The graph information acquisition module 704 is used to convert the routes and transportation hubs in the database into the graph structure in the graph neural network according to the starting point and the end point, and generate the graph structure and graph structure information of the target area; the graph structure uses the interchange hubs between roads as nodes, and the roads between the interchange hubs as edges; the graph structure information includes node feature information and edge feature information; the node feature information includes the traffic restriction information of the node, and the edge feature information includes the traffic safety information of the road; the traffic safety information includes whether the road meets the traffic requirements of large-scale transport vehicles, or the safety factor and precautions for the passage of large-scale transport vehicles.

[0105] The prediction module 706 is used to input the starting point, the end point, the vehicle information, the graph structure of the target area and the graph structure information into a pre-trained prediction model, and output the optimal transportation path of the large-scale transport vehicle through the prediction model. The prediction model is a neural network model.

[0106] In one embodiment, the traffic restriction information of the node includes: vehicle height restriction, vehicle length restriction, driving width restriction, and maximum load restriction.

[0107] In one of the embodiments, the traffic safety information also includes the traffic safety information of large-scale transport vehicles on the road; the training loss of the prediction model includes three basic constraints: nodes and paths that exceed the traffic limit are not within the path range; minimization of the planned path length, and maximization of the traffic safety of the traffic path.

[0108] In one of the embodiments, an updating module is further included, which is used to train the prediction model according to the graph structure and graph structure information of the updated area when the graph structure and graph structure information of the updated area are obtained, so as to update the prediction model.

[0109] In one embodiment, the prediction model uses a graph neural convolutional network; the convolutional layer of the graph neural convolutional network can consider the weight influence of nodes with edge connections when processing the node information depth, and the calculation formula is:

[0110] ;

[0111] in, and is the fully connected layer (MLP) of the neural network; x i (k-1) and x i (k) They are k -1st floor and next floor k Feature vector of layer data points; is the edge connection eigenvector, x j (k-1) is the corresponding feature vector of nearby data points.

[0112] In one embodiment, the graph neural convolution network adopts a graph convolution aggregation method, and its node information depth aggregation relationship is:

[0113] ;

[0114] in,H (l) and H (l+1) They are the node information representations before and after aggregation respectively; represents the activation function; Represents the adjacency matrix between nodes; yes The degree matrix of W (l) yes l Layer and l +1 Neural network training parameters between layers.

[0115] Each module in the above-mentioned large-scale transport vehicle transportation route planning system based on AI algorithm can be implemented in whole or in part through software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0116] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used for predicting models. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a transportation path planning method for large-scale transportation vehicles based on an AI algorithm is implemented.

[0117] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0118] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the method for planning a transportation path for large-scale transport vehicles based on an AI algorithm in the above-mentioned embodiments are implemented.

[0119] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for planning a transportation path for large-scale transport vehicles based on an AI algorithm according to the above-mentioned embodiments is implemented.

[0120] In one embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned methods for planning transport routes for large-scale transport vehicles based on AI algorithms.

[0121] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The above-mentioned computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0122] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0123] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for transport path planning of large-scale transport vehicles based on AI algorithm, characterized in that: The method comprises: Build a database of route traffic information and transportation hub information in the region; Obtain the starting point, end point and vehicle information of large-scale transport vehicles; According to the starting point and the end point, the routes and transportation hubs in the database are converted into a graph structure in a graph neural network to generate a graph structure and graph structure information of the target area; the graph structure uses the interconnection hubs between the roads as nodes and the roads between the interconnection hubs as edges; the graph structure information includes node feature information and edge feature information; the node feature information includes the traffic restriction information of the node, and the edge feature information includes the traffic safety information of the road; the traffic safety information includes whether the road meets the traffic requirements of large-scale transport vehicles, or the safety factor and precautions for the passage of large-scale transport vehicles; Input the starting point, the end point, the vehicle information, the graph structure of the target area and the graph structure information into a pre-trained prediction model, and output the optimal transportation path of the large-scale transportation vehicle through the prediction model, wherein the prediction model is a neural network model; The prediction model adopts a graph neural convolutional network; the convolutional layer of the graph neural convolutional network can consider the weight influence of nodes with edge connections when processing the node information depth, and its calculation formula is: ； in and is the MLP of the neural network; x i (k-1) and x i (k) They are k -1st floor and next floor k Feature vector of layer data points; is the edge connection eigenvector, x j (k-1) is the corresponding feature vector of nearby data points.

2. The method according to claim 1, characterized in that The traffic restriction information of the node includes: vehicle height restriction, vehicle length restriction, driving width restriction and maximum load restriction.

3. The method according to claim 2, characterized in that The traffic safety information also includes the traffic safety information of large-scale transport vehicles on the road; the training loss of the prediction model includes three basic constraints: nodes and paths that exceed the traffic limit are not within the path range; minimization of the planned path length, and maximization of the traffic safety of the traffic path.

4. The method according to claim 1, characterized in that: The method further comprises: When the graph structure and graph structure information of the update area are obtained, the prediction model is trained according to the graph structure and graph structure information of the update area to update the prediction model.

5. The method according to claim 1, characterized in that The graph neural convolution network adopts the graph convolution aggregation method, and its node information deep aggregation relationship is: ； in, H (l) and H (l+1) They are the node information representations before and after aggregation respectively; represents the activation function, Represents the adjacency matrix between nodes; yes The degree matrix of W (l) yes l Layer and l +1 Neural network training parameters between layers.

6. A large-scale transport vehicle transportation route planning system based on AI algorithm, characterized in that: The system comprises: Construction module, used to build a database of route traffic information and transportation hub information in the region; Planning acquisition module, used to obtain the starting point, end point and vehicle information of large-scale transport vehicles; A graph information acquisition module is used to convert the routes and transportation hubs in the database into a graph structure in a graph neural network according to the starting point and the end point, and generate a graph structure and graph structure information of the target area; the graph structure uses the interconnection hubs between the roads as nodes, and the roads between the interconnection hubs as edges; the graph structure information includes node feature information and edge feature information; the node feature information includes the traffic restriction information of the node, and the edge feature information includes the traffic safety information of the road; the traffic safety information includes whether the road meets the traffic requirements of large-scale transport vehicles, or the safety factor and precautions for the passage of large-scale transport vehicles; A prediction module, used to input the starting point, the end point, the vehicle information, the graph structure of the target area and the graph structure information into a pre-trained prediction model, and output the optimal transportation path of the large-scale transportation vehicle through the prediction model, wherein the prediction model is a neural network model; The prediction model adopts a graph neural convolutional network; the convolutional layer of the graph neural convolutional network can consider the weight influence of nodes with edge connections when processing the node information depth, and its calculation formula is: ； in, and is the MLP of the neural network; x i (k-1) and x i (k) They are k -1st floor and next floor k Feature vector of layer data points; is the edge connection eigenvector, x j (k-1) is the corresponding feature vector of nearby data points.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images