A data processing method and device, electronic equipment and storage medium

Abstract

The application discloses a data processing method and device, electronic equipment and storage medium, wherein the method comprises: determining a traffic event graph of a target region at a target time, wherein a node of the traffic event graph represents a traffic event entity, an edge of the traffic event graph represents a corresponding relationship between two traffic event entities, and the traffic event entity comprises at least one of a traffic event occurring in the target region at the target time, position information of the traffic event, and a cause of the traffic event; inputting the traffic event graph into a target decision model to obtain a target traffic control strategy for the target region at the target time; wherein the target decision model is determined according to a historical traffic event graph of the target region at a first time and a historical traffic control strategy for the target region corresponding to the historical traffic event graph, and the first time comprises at least one time before the target time.

Description

Technical Field

[0001] This application belongs to the field of intelligent transportation technology, specifically relating to a data processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] Relevant traffic problem analysis methods generally involve experts abstracting traffic management knowledge into tags such as traffic problems, causes, and strategies based on a case database and storing them in a relational database. Database operations are then used to identify different types of problems. However, traffic congestion and overflow are complex phenomena resulting from the interaction of multiple issues. Relational data struggles to represent the relationships between these issues, making it difficult to provide comprehensive recommendations when multiple problems influence each other. Secondly, human experience is limited, making it difficult to extend to existing scenarios. Furthermore, it can only analyze problems within a single time period, while traffic problems are dynamic and time-varying; analyzing only one time dimension makes it difficult to provide reasonable strategies for traffic management. Summary of the Invention

[0003] The purpose of this application is to provide a data processing method, apparatus, electronic device, and storage medium that can improve the comprehensiveness and accuracy of traffic incident identification and cause analysis, and enable automatic recommendation of traffic control strategies.

[0004] To solve the above-mentioned technical problems, this application is implemented as follows:

[0005] In a first aspect, embodiments of this application provide a data processing method, the method comprising: determining a traffic event map of a target area at a target time, wherein nodes of the traffic event map represent traffic event entities, edges of the traffic event map represent the correspondence between two traffic event entities, and the traffic event entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event; inputting the traffic event map into a target decision model to obtain a target traffic control strategy for the target area at the target time; wherein the target decision model is determined based on a historical traffic event map of the target area at a first time and a historical traffic control strategy for the target area corresponding to the historical traffic event map, the first time including at least one time prior to the target time.

[0006] Secondly, embodiments of this application provide a data processing apparatus, comprising: a determining module, configured to determine a traffic event map of a target area at a target time, wherein nodes of the traffic event map represent traffic event entities, edges of the traffic event map represent the correspondence between two traffic event entities, and the traffic event entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event; and a processing module, configured to input the traffic event map into a target decision model to obtain a target traffic control strategy for the target area at the target time; wherein the target decision model is determined based on a historical traffic event map of the target area at a first time and a historical traffic control strategy for the target area corresponding to the historical traffic event map, the first time including at least one time prior to the target time.

[0007] Thirdly, embodiments of this application provide an electronic device including a processor and a memory, the memory storing programs or instructions executable on the processor, the programs or instructions, when executed by the processor, implementing the steps of the method described in the first aspect.

[0008] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0009] In this embodiment, a traffic event map of a target area at a target time is determined. Nodes in the traffic event map represent traffic event entities, and edges represent the correspondence between two traffic event entities. Each traffic event entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event. The traffic event map is input into a target decision model to obtain a target traffic control strategy for the target area at the target time. The target decision model is determined based on a historical traffic event map of the target area at a first time and corresponding historical traffic control strategies for the target area. The first time includes at least one time prior to the target time. By combining the structural and temporal characteristics of the traffic event map and utilizing a temporal graph neural network model, the comprehensiveness and accuracy of traffic event identification and causal analysis are improved, and automatic recommendation of traffic control strategies is achieved. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating a data processing method provided in an embodiment of this application.

[0012] Figure 2 This is a schematic diagram of metadata information for a traffic event map provided in an embodiment of this application.

[0013] Figure 3 This is a network structure diagram of a target decision model provided in an embodiment of this application.

[0014] Figure 4 This is a schematic diagram of the structure of a data processing device provided in an embodiment of this application.

[0015] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0016] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0017] Figure 1 This illustration shows a flowchart of a data processing method provided in an embodiment of this application. This method can be executed by an electronic device, such as a terminal device or a server device. In other words, the method can be executed by software or hardware installed on the terminal device or server device. The server device includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. Figure 1 As shown, the method may include the following steps.

[0018] S101: Determine the traffic event map for the target area at the target time.

[0019] A knowledge graph, known in the library and information science field as knowledge domain visualization or knowledge domain mapping map, is a series of different graphics that display the development process and structural relationships of knowledge. It uses visualization technology to describe knowledge resources and their carriers, and to mine, analyze, construct, draw and display knowledge and the interrelationships between them.

[0020] In this embodiment of the application, the traffic event graph represents a mapping map of traffic events in the field of intelligent transportation. Specifically, the nodes of the traffic event graph represent traffic event entities, and the edges of the traffic event graph represent the correspondence between two traffic event entities. The traffic event entity includes at least one of the following: a traffic event that occurred in the target area at the target time, the location information of the traffic event, and the cause of the traffic event.

[0021] S102: Input the traffic event map into the target decision model to obtain the target traffic control strategy for the target area at the target time.

[0022] The target decision model is determined based on the historical traffic event map of the target area at a first time and the corresponding historical traffic control strategy for the target area, wherein the first time includes at least one time prior to the target time.

[0023] The first time is at least one time prior to the target time, corresponding to the historical traffic event map being time-series data. The target decision model learns information from the time-series map data and the corresponding label relationship (i.e., the historical traffic control strategies for the target area corresponding to the historical traffic event map), and finally determines the learned information in the form of model parameters. The historical traffic control strategy is used to process each historical traffic event in the historical traffic event map, such as signal control optimization strategy and organization optimization strategy. The signal control optimization strategy includes optimization cycle and optimization of straight-through green light ratio. The optimization of traffic control strategies includes: optimizing the left-turn green light ratio, optimizing the maximum and minimum green light ratios, optimizing green light intervals, optimizing phase differences, optimizing phase structure, optimizing phase sequence, upstream and downstream coordination control, dedicated left-turn control, right-turn control, adding traffic light control, and increasing pedestrian clearance time. Specifically, optimizing phase structure includes symmetrical release, overlapping release, single-entry release, and repeated release. Optimizing phase sequence includes staggering conflicting phases and releasing traffic in the same phase. Organizational optimization strategies include balancing the number of lanes, adding dedicated lanes, adding guide lines, adding left-turn waiting areas, adding straight-ahead waiting areas, adding safety islands, prioritizing U-turns, postponing pedestrian crossings, and setting speed limits for road sections. This step inputs the traffic event map of the target area at the target time into the target decision model to obtain the target traffic control strategy for the target area at the target time.

[0024] This application provides a data processing method that determines a traffic event map of a target area at a target time. Nodes in the traffic event map represent traffic event entities, and edges represent the correspondence between two traffic event entities. Each traffic event entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event. The traffic event map is input into a target decision model to obtain a target traffic control strategy for the target area at the target time. The target decision model is determined based on a historical traffic event map of the target area at a first time and corresponding historical traffic control strategies for the target area. The first time includes at least one time prior to the target time. By combining the structural and temporal characteristics of the traffic event map and utilizing a temporal graph neural network model, the method improves the comprehensiveness and accuracy of traffic event identification and causal analysis, and enables automatic recommendation of traffic control strategies.

[0025] In one implementation, step S101 includes: when there is a corresponding relationship between the two traffic event entities, determining one of the traffic event entities as the head entity and the other traffic event entity as the tail entity; using the head entity as the starting node of the edge of the traffic event graph and the tail entity as the terminal node of the edge of the traffic event graph.

[0026] In one implementation, the head entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event; the tail entity includes at least one traffic event occurring in the target area at the target time.

[0027] In one implementation, the location information includes at least one of an intersection, a road segment, and a traffic flow direction; when the head entity includes at least one of an intersection, a road segment, and a traffic flow direction, the tail entity also includes at least one of an intersection, a road segment, and a traffic flow direction.

[0028] In this embodiment of the application, the traffic time map displays the correlation between different intersections and different time points, representing different dimensions of problems, thus constructing a multi-dimensional traffic problem map of "time-space-multi-model". Specifically, Figure 2 This illustration shows metadata information for a traffic event map provided in an embodiment of this application. For example... Figure 2As shown, this application embodiment can abstract 5 types of traffic event entities and 8 types of relationships. Traffic event entities can include intersections, road segments, flow directions, traffic problems, and causes of traffic problems; relationships include intersection_connecting_road segments, road segments_connecting_intersections, road segments_containing_flow directions, intersection_exists_problem, road segments_exists_problem, flow directions_exists_problem, cause_leads to_problem, and problem_leads to_problem.

[0029] In one implementation, prior to the aforementioned step S101, the method further includes: determining the plurality of traffic event entities based on preset road network information and multimedia information collected by road network equipment.

[0030] For example, the entity relationship information of intersections, road segments and flow directions is extracted by the entity relationship extraction module. The module obtains dynamic and static basic data from road network information and equipment detection data, and extracts entity relationships at 15-minute intervals to construct a time-series traffic event map (TKG). The TKG consists of multiple time-series static traffic event maps: TKG = {SKG1, SKG2, SKG3, ..., SKGt}, where SKGt = {E, R, Ft}, E and R are the sets of entity relationships, and Ft is the time-series relationship set {eh, r, et, t} at time t, indicating that the head entity eh and the tail entity et have a relationship r at time t. Where entity E = {I, R, F, P, C} represents intersection entity, road segment entity, flow direction entity, traffic problem entity, and traffic problem cause entity, respectively. Relation R = {IR, RI, RF, IP, RP, FP, CP, PP} represents intersection_connecting_road segment, road segment_connecting_intersection, road segment_containing_flow direction, intersection_exists_problem, road segment_exists_problem, flow direction_exists_problem, cause_leads_to_problem, and problem_leads_to_problem relationship, respectively. At time t, there are Nti intersection entities {I1,I2,I3,……,INti}, Ntr road segment entities {R1,R2,R3,……,RNtr}, Ntf flow entities {F1,F2,F3,……,FNtf}, Ntp problem entities {P1,P2,P3,……,PNtp}, and Ntc cause entities {C1,C2,C3,……,CNtc}. The synchronization relationships are Ntir, Ntri, Ntrf, Ntip, Ntrp, Ntfp, Ntcp, and Ntpp.

[0031] This allows for the design of various attributes for each traffic event entity and relationship. For example, intersection entities can include attributes such as intersection time number, intersection number, intersection name, intersection type, intersection latitude and longitude, whether the intersection is controlled, intersection service level, number of lanes, timestamp, and entity class name; road segment entities can include attributes such as road segment time number, road segment number, road segment name, road segment type, road segment direction number, road segment direction name, road segment length, number of lanes entering the road segment, whether the road segment is closed, road segment straight-left unevenness coefficient, road straight unevenness coefficient, timestamp, entity class name, and number of lanes exiting the road segment; flow direction entities include turning... The traffic problem entity includes attributes such as traffic problem number, traffic problem name, traffic problem description, and entity class name; the traffic problem cause entity includes attributes such as traffic problem cause number, traffic problem cause name, traffic problem cause description, and entity class name; the intersection connection segment relationship includes intersection number, segment number, and entity class name. The attributes include: class name, intersection-connected road segment time number, intersection-connected road segment number, timestamp, and relationship name; the road segment-to-intersection relationship includes: road segment number, intersection number, entity class name, road segment-to-intersection time number, road segment-to-intersection number, timestamp, and relationship name; the road segment-containing flow direction relationship includes: road segment number, flow rate number, road segment-containing flow direction time number, road segment-containing flow direction number, and entity class name; the intersection-occurring problem relationship includes: intersection number, problem number, intersection-occurring problem number, entity class name, and intersection-occurring problem name; the road segment-issuing problem relationship includes: The relationship between a problem and an event includes attributes such as road segment number, problem number, problem number occurring on the road segment, entity class name, and name of the problem occurring at the intersection; the relationship between an event and a problem includes attributes such as event number, problem number, intersection number where the event occurred, entity class name, and name of the event; the relationship between a cause and a problem includes attributes such as cause number, problem number, cause leading to problem number, part number of cause leading to problem, entity class name, and name of the cause leading to event; the relationship between a problem and a problem includes attributes such as problem number 1, problem number 2, problem leading to problem number, entity class name, and name of the problem leading to problem.

[0032] In addition, in this embodiment, traffic problems, causal entities, and relationships can be identified through a pre-set traffic problem diagnostic model to establish a tag library. Problem tags are categorized into efficiency, safety, and violation issues. Efficiency problems include uneven saturation of opposing traffic, large delays in straight-ahead signaling, large delays in left-turn signaling, numerous stops for straight-ahead and left-turn traffic, long queues for straight-ahead, left-turn, and right-turn traffic, downstream overflow, low vehicle speeds, excessive waiting times for crossings, interference between motorized and non-motorized traffic, and overflow in certain sections. Safety problems include merging conflicts, intersection conflicts, diverging conflicts, left-turn weaving conflicts, right-turn weaving conflicts, weaving conflicts in widened sections, motorized-motorized conflicts, motorized-non-motorized conflicts, large speed fluctuations, frequent cutting in, frequent lane changes, and motorized-pedestrian conflicts. Violation issues include pedestrians running red lights and vehicles parking illegally. Causal tags are categorized into signal control, traffic organization and channelization, travel demand, and traffic... Incidents and violations, signal control issues such as unreasonable cycle times, unreasonable straight-ahead green light ratios, unreasonable left-turn green light ratios, unreasonable phase sequences, unreasonable phase structures, unreasonable green light intervals, unreasonable minimum green light ratios, unreasonable maximum green light ratios, unreasonable phase differences, lack of coordination between upstream and downstream traffic, uncontrolled right turns, lack of dedicated left-turn control, uncontrolled intersections, unreasonable pedestrian clearance times, ineffective green wave effects, and uncontrolled pedestrian crossings; and organization and channelization issues such as unreasonable lane hierarchy differences, lack of dedicated lanes, mismatched number of merging lanes, excessive changes in the number of lanes in widened sections, excessively long pedestrian crossings, lack of guide lines, lack of waiting areas, unreasonable number of lanes, mixed-traffic lanes adjacent to dedicated lanes, multiple right-turn lanes, multiple left-turn lanes, and large intersection areas. T-junctions, excessively short downstream road segments, and relatively short road segments are among the traffic issues that affect travel demand, including high traffic volume at intersections, high volume of straight-through vehicles, high volume of left-turning vehicles, high volume of right-turning vehicles, high volume of U-turn vehicles, high pedestrian traffic, a high proportion of large vehicles, a high proportion of buses, uneven traffic flow in the same direction, uneven traffic flow in opposite directions, holiday travel, peak travel, off-peak travel, high volume of non-motorized vehicles, and numerous points of interest. Traffic incidents include speed limits, road closures, traffic accidents, and severe weather. Traffic violations include speeding, crossing solid lines, illegal parking, pedestrians jaywalking, and cutting in line. To address these traffic issues, as mentioned earlier, traffic control strategies are categorized into signal control optimization and organizational optimization. Signal control optimization can include optimization cycles and optimization of straight-through traffic. The tag library includes optimizations such as green light ratio, left-turn green light ratio, maximum green light ratio, minimum green light ratio, green light interval, phase difference, phase structure, phase sequence, upstream and downstream coordination control, dedicated left-turn control, right-turn control, adding traffic light control, and increasing pedestrian clearance time. Optimizing phase structure specifically includes symmetrical clearance, overlapping clearance, single-entry clearance, and repeated clearance. Optimizing phase sequence includes staggering conflicting phases and releasing traffic in the same phase. Organizational optimization includes balancing the number of lanes, adding dedicated lanes, adding guide lines, adding left-turn waiting areas, adding straight-ahead waiting areas, adding safety islands, prioritizing U-turns, placing pedestrian crossings behind, and setting speed limits for road sections. This tag library can be continuously improved and accumulated in traffic scenarios, and this application does not impose specific restrictions.

[0033] In one implementation, there are multiple first time periods, and the target decision model includes a graph convolutional network layer and a recurrent neural network layer. Before inputting the traffic event map into the target decision model to obtain the target traffic control strategy for the target area at the target time, the method further includes: inputting multiple historical traffic event maps of the target area at multiple first time periods into a graph convolutional network layer to be trained to obtain multiple first traffic event codes for the target area at multiple first time periods; inputting the multiple first traffic event codes into a recurrent neural network layer to be trained in chronological order to obtain multiple predicted traffic control strategies for the target area at multiple first time periods; and determining the target decision model based on the multiple predicted traffic control strategies and the multiple historical traffic control strategies.

[0034] Figure 3 The diagram illustrates a network structure of a target decision model provided in an embodiment of this application. Specifically, as shown... Figure 3 As shown in this embodiment, a Relational Graph Convolutional Network (R-GCN) is first used to reduce the dimensionality of multiple traffic event graphs (Temp-1, Temp-2, ..., Temp-n) at the first time point into multiple first traffic event codes, which is the process of graph embedding. Then, each first traffic event code is sequentially input into multiple gated recurrent units (GRUs) (a type of recurrent neural network) in chronological order to obtain the corresponding temporal features.

[0035] Compared to traditional spatial feature extractors like Convolutional Neural Networks (CNNs), R-GCN can handle non-Euclidean structure data such as road network structures, making it suitable for global feature extraction. Compared to basic Graph Convolutional Networks (GCNs), R-GCN can solve graph embedding of multiple relationships in heterogeneous graphs. Furthermore, embodiments of this application use RNNs to capture the dynamics of sequences in traffic event states. However, RNNs have limitations in long-term prediction due to gradient vanishing and exploding when performing long products. LSTM and GRU models are variants of RNNs. GRU addresses the shortcomings of RNNs by using a gating mechanism to maintain long-term memory. GRU is simpler, easier to train, has fewer parameters, and trains faster than LSTM.

[0036] In this embodiment, the traffic strategy recommendation problem can be abstracted as a graph multi-classification problem. A function f is used to classify traffic conditions based on known information.

[0037] {y1,y2…y l}=f θ (X,G)

[0038] Among them, y l Let θ represent the number of classes and θ represent all learnable parameters in the model.

[0039] The model utilizes spectral convolution operations between graph nodes and their first-order neighbor domains to capture the spatial features of the graph. Similar to CNNs, the hierarchical propagation rule also applies to stacking multiple networks. The node representation after a multi-layer GCN model can be defined as:

[0040]

[0041] in This refers to the l-th level node representation of node i. This refers to the l-th level node representation of all neighboring nodes of node j. c represents the set of neighboring nodes of node i whose relation is r; i,r It is a regularization constant, where c i,r The value is W r (l) It is a linear transformation function that transforms neighboring nodes of the same type of edge using a parameter matrix W. t (l) Perform a transformation; for neighboring nodes connected by different edge types, perform a linear transformation, W t (l) The number of edges is the same as the number of edge types, and σ(·) represents a non-linear activation function similar to ReLU.

[0042] R-GCN can encode the topology and characteristics of traffic flows by determining the topological relationships between central traffic flows and surrounding traffic flows. In summary, the R-GCN model learns spatial dependencies.

[0043] GRU has a reset gate and an update gate. The reset gate is used to determine the remaining proportion of the new input, and the update gate is used to capture the previous input saved to the current time step. For a given time step t, the update gate r t and reset door u t The calculation is as follows:

[0044] u t =σ(W (u) *[x t +h t-1 ]+b u)

[0045] r t =σ(W (r) *[x t +h t-1 ]+b r )

[0046] Where, x t It is the input for time step t, h t-1 It is the hidden state of the previous time step. Finally, through the old state h t-1 and the new candidate hidden state h t By performing element-wise multiplication between the nodes, we obtain the final update equation for GRU as follows:

[0047] h t ′=tanh(W (c) [x t ,r t *h t-1 ]+b c )

[0048] h t =u t *h t-1 +(1-u t )*h t ′

[0049] The GRU model can capture traffic information at the current time step while preserving historical traffic trends. Through the GRU layer, temporal features are extracted from multiple historical traffic events at the first time step, and then connected to a classifier to predict traffic events and recommend traffic strategies.

[0050] In one implementation, the step of inputting the traffic event map into the target decision model to obtain the target traffic control strategy for the target area at the target time includes: inputting the traffic event map into the target decision model to obtain the probability of multiple pre-selected traffic control strategies for the target area at the target time; and determining the top n pre-selected traffic control strategies with the highest probability among the multiple pre-selected traffic control strategies as the target traffic control strategy, where n is an integer greater than or equal to 1.

[0051] The traffic problem in this embodiment is a multi-label classification task. After the target decision model is trained, a classifier is connected after the recurrent neural network to obtain the probability of each category. This step can select the categories with the highest probabilities as the final target control strategy, and the control strategy with higher probability also has higher priority.

[0052] In relevant datasets, there may be imbalanced data distribution, with varying levels of difficulty in identifying different traffic problems. While easily classified samples may have lower loss, their large number can significantly contribute to the overall loss, preventing the model from converging to a satisfactory result. In such cases, focal loss can be used to reduce the weight of easily classified samples, allowing training to focus more on more meaningful samples. Unlike the traditional cross-entropy loss function, focal loss is calculated as follows:

[0053] CE(p t ) = -log(p t )

[0054] FL(p t )=-(1-p t ) γ log(p t )

[0055] Compared to cross-entropy, focal loss adds a factor (1-p). t ) γ When γ = 0, it is the same as cross-entropy; when γ > 0, it can reduce the relative loss of well-classified samples and focus more attention on misclassified examples that are difficult to classify.

[0056] Figure 4 This application provides a schematic diagram of the structure of a data processing apparatus according to an embodiment of the present application. Figure 4 As shown, the device 400 includes a determination module 410 and a processing module 420.

[0057] The determining module 410 is used to determine a traffic event map of a target area at a target time. Nodes in the traffic event map represent traffic event entities, and edges in the traffic event map represent the correspondence between two traffic event entities. Each traffic event entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event. The processing module 420 is used to input the traffic event map into a target decision model to obtain a target traffic control strategy for the target area at the target time. The target decision model is determined based on a historical traffic event map of the target area at a first time and a historical traffic control strategy for the target area corresponding to the historical traffic event map. The first time includes at least one time prior to the target time.

[0058] In one implementation, the determining module 410 is configured to: when there is a corresponding relationship between two traffic event entities, determine one of the traffic event entities as the head entity and the other traffic event entity as the tail entity; use the head entity as the starting node of the edge of the traffic event graph and the tail entity as the terminal node of the edge of the traffic event graph.

[0059] In one implementation, the head entity includes at least one of the following: a traffic event occurring in the target area at the target time, location information of the traffic event, and the cause of the traffic event; the tail entity includes at least one traffic event occurring in the target area at the target time.

[0060] In one implementation, the apparatus further includes a training module, configured to: input multiple historical traffic event maps of the target area at multiple first times into a graph convolutional network layer to be trained, to obtain multiple first traffic event codes of the target area at multiple first times; input the multiple first traffic event codes sequentially into a recurrent neural network layer to be trained in chronological order, to obtain multiple predicted traffic control strategies for the target area at multiple first times; and determine the target decision model based on the multiple predicted traffic control strategies and the multiple historical traffic control strategies.

[0061] In one implementation, the determining module 410 is further configured to: determine the plurality of traffic event entities based on preset road network information and multimedia information collected by road network equipment.

[0062] In one implementation, the processing module 420 is configured to: input the traffic event map into the target decision model to obtain the probability of the target time for multiple pre-selected traffic control strategies for the target area; and determine the top n pre-selected traffic control strategies with the highest probability among the multiple pre-selected traffic control strategies as the target traffic control strategy, where n is an integer greater than or equal to 1.

[0063] The data processing apparatus provided in this application embodiment can execute the data processing method described in at least one embodiment of the preceding method embodiment, and can achieve the same technical effect as the preceding method embodiment, which will not be repeated here.

[0064] Figure 5This diagram illustrates the hardware structure of an electronic device implementing the embodiments of this application. Referring to the diagram, at the hardware level, the electronic device includes a processor and optionally, an internal bus, a network interface, and a memory. The memory may include RAM, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk storage device. Of course, the electronic device may also include other hardware required for other services.

[0065] The processor, network interface, and memory can be interconnected via an internal bus, which can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be categorized as an address bus, data bus, control bus, etc. For ease of illustration, only a single bidirectional arrow is used in this diagram, but this does not imply that there is only one bus or one type of bus.

[0066] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0067] The processor reads the corresponding computer program from non-volatile memory into memory and then runs it, forming a device for locating the target user at the logical level. The processor executes the program stored in memory and specifically performs each step of the data processing method described in at least one embodiment of the above-described methods.

[0068] The methods disclosed in the embodiments shown in the flowcharts of this application can be applied to a processor or implemented by a processor. The processor may be an integrated circuit with signal processing capabilities. During implementation, each step of the above methods can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0069] The electronic device can also perform the data processing method described in at least one embodiment of the method embodiments above, and can achieve the same technical effect as the method embodiments above, which will not be repeated here.

[0070] Of course, in addition to software implementation, the electronic device of this application does not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. In other words, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0071] This application also proposes a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform the steps of the data processing method described in at least one embodiment of the above-described method embodiments.

[0072] The computer-readable storage medium includes read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, etc.

[0073] Furthermore, this application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, which, when executed by a computer, implement the steps of at least one embodiment of the above-described method embodiments.

[0074] In summary, the above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

[0075] The systems, devices, modules, or units described in the above embodiments can be implemented by a computer or by a product with a certain function. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0076] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0077] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0078] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

Claims

1. A data processing method, characterized in that, The method includes: A traffic event map is determined in a target area at a target time. The nodes of the traffic event map represent traffic event entities, and the edges of the traffic event map represent the correspondence between two traffic event entities. The traffic event entity includes: a traffic event that occurred in the target area at the target time, the location information of the traffic event, and the cause of the traffic event. Input the traffic event map into the target decision model to obtain the target traffic control strategy for the target area at the target time. The target decision model is trained using the following method: Input the historical traffic event maps of the target area at multiple first times prior to the target time into the graph convolutional network layer to be trained to obtain multiple first traffic event codes of the target area at multiple first times; Multiple first traffic event codes are sequentially input into the recurrent neural network layer to be trained in chronological order to obtain multiple predicted traffic control strategies for the target area at multiple first times. Based on multiple predicted traffic control strategies and multiple historical traffic control strategies, a target decision model is determined.

2. The method according to claim 1, characterized in that, The determination of the traffic event map in the target time and target area includes: When there is a corresponding relationship between two traffic event entities, one of the traffic event entities is identified as the head entity and the other traffic event entity is identified as the tail entity. The head entity is used as the starting node of the edge in the traffic event graph, and the tail entity is used as the ending node of the edge in the traffic event graph.

3. The method according to claim 2, characterized in that, The head entity includes at least one of the following: a traffic event that occurred in the target area at the target time, location information of the traffic event, and the cause of the traffic event; the tail entity includes at least one traffic event that occurred in the target area at the target time.

4. The method according to claim 3, characterized in that, The location information includes at least one of an intersection, a road segment, and a traffic flow direction; when the head entity includes at least one of an intersection, a road segment, and a traffic flow direction, the tail entity also includes at least one of an intersection, a road segment, and a traffic flow direction.

5. The method according to claim 4, characterized in that, Before determining the traffic event map for the target time and target area, the following steps are also included: Based on the preset road network information and the multimedia information collected by the road network equipment, multiple traffic event entities are identified.

6. The method according to claim 1, characterized in that, The traffic event map is input into the target decision model to obtain the target traffic control strategy for the target area at the target time, including: The traffic event map is input into the target decision model to obtain the probability of multiple pre-selected traffic control strategies for the target area at the target time. Among the multiple pre-selected traffic control strategies, the one with the highest probability is selected first. A pre-selected traffic control strategy is determined as the target traffic control strategy, wherein... It is an integer greater than or equal to 1.

7. A data processing apparatus, characterized in that, The device includes: The determination module is used to determine a traffic event map in a target area at a target time. The nodes of the traffic event map represent traffic event entities, and the edges of the traffic event map represent the correspondence between two traffic event entities. The traffic event entity includes: a traffic event that occurred in the target area at the target time, the location information of the traffic event, and the cause of the traffic event. The training module is used to input multiple historical traffic event maps of the target area at multiple first times prior to the target time into a graph convolutional network layer to be trained, thereby obtaining multiple first traffic event codes for the target area at multiple first times; input the multiple first traffic event codes into a recurrent neural network layer to be trained in chronological order, thereby obtaining multiple predicted traffic control strategies for the target area at multiple first times; and determine a target decision model based on the multiple predicted traffic control strategies and the multiple historical traffic control strategies. The processing module is used to input the traffic event map into the target decision model to obtain the target traffic control strategy for the target area at the target time.

8. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method as described in any one of claims 1-6.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1-6.

Images