Traffic scheduling method, device, equipment, medium and program product

Abstract

The invention relates to a traffic scheduling method, device and equipment, a medium and a program product. The method comprises the steps that real-time traffic data of a target area is acquired, the real-time traffic data can be further processed through a pre-trained deep learning model, a traffic prediction result can represent the future traffic condition of the target area, and a traffic scheduling strategy is formulated and executed based on the obtained traffic prediction result. The deep learning model is trained by using the traffic data in advance, so that the deep learning model can automatically learn traffic modes and rules, a high-precision and real-time traffic prediction result is provided according to the real-time traffic data, the traffic scheduling difficulty is reduced, effective scheduling can be performed by means of the traffic prediction result, traffic congestion is relieved, and the traffic scheduling efficiency is improved. The traffic condition is improved.

Description

Technical Field

[0001] The present invention relates to the field of traffic dispatching, and in particular to a traffic dispatching method, device, equipment, medium and program product. Background Art

[0002] With the advancement of technology and the development of the automobile industry, the traffic volume in large cities, ports and transportation hubs is increasing, and traffic management and dispatching are facing increasingly complex challenges. Especially in complex scenarios with traffic nodes such as traffic lights and variable lanes, how to conduct reasonable dispatch to alleviate traffic conditions is a major problem.

[0003] At present, traffic dispatch solutions in related technologies usually rely on rule-driven algorithms. This approach has poor data processing capabilities and is only applicable to simple road scenarios. It is difficult to cope with increasingly complex traffic conditions, resulting in increasingly serious problems such as traffic congestion and frequent accidents. Summary of the invention

[0004] One of the purposes of the present invention is to provide a traffic dispatching method, device, equipment, medium and program product to solve the problem of being unable to perform reasonable dispatch according to traffic conditions.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A traffic dispatching method, comprising:

[0007] Obtain real-time traffic data in the target area;

[0008] Processing the real-time traffic data through a pre-trained deep learning model to obtain a traffic prediction result, wherein the traffic prediction result represents the future traffic conditions of the target area;

[0009] A traffic dispatching strategy is formulated and executed based on the traffic prediction results. The traffic dispatching strategy includes dispatching traffic nodes and / or displaying traffic guidance information. The dispatching traffic nodes includes adjusting the duration of traffic lights in the target area and / or regulating roads in the target area.

[0010] Further, a traffic dispatching strategy is formulated and executed based on the traffic prediction result, including:

[0011] Initialize the first generation of individuals based on the genetic algorithm, the genes of the first generation of individuals represent the behavior of the dispatching traffic node;

[0012] Determining a fitness function according to the traffic prediction result;

[0013] Iterating the first generation of individuals by changing the value of the gene, and calculating the fitness value of each individual according to the fitness function;

[0014] Selecting a target individual according to the fitness value until an iteration termination condition is reached;

[0015] After the iteration is terminated, a traffic dispatch strategy is formulated based on the genes of the selected target individuals.

[0016] Further, determining a fitness function according to the traffic prediction result includes:

[0017] Selecting at least one indicator from traffic flow, congestion index and accident risk according to the traffic prediction result;

[0018] Construct a fitness function based on the selected indicators.

[0019] Furthermore, the real-time traffic data includes point cloud features, image features, and indication information of traffic nodes in the target area, and the processing of the real-time traffic data by the pre-trained deep learning model includes:

[0020] Mapping the point cloud features and the image features to the three-dimensional space corresponding to the target area respectively, and determining the semantic information of the three-dimensional space according to the mapping results;

[0021] The indication information and the semantic information are input into a pre-trained deep learning model to obtain a traffic prediction result output by the deep learning model.

[0022] Furthermore, the pre-trained deep learning model is obtained in the following way:

[0023] Get historical traffic data;

[0024] Constructing a training set and a validation set based on the historical traffic data;

[0025] Training an initial model based on the data in the training set, wherein the initial model is a deep learning model based on a bird's-eye view fusion architecture;

[0026] The parameters of the trained initial model are adjusted based on the data in the validation set to obtain a pre-trained deep learning model.

[0027] Furthermore, traffic guidance information is displayed, including:

[0028] Displaying the traffic guidance information on a display screen within the target area;

[0029] and / or,

[0030] The traffic guidance information is pushed to the terminal device used by the user.

[0031] A traffic dispatching device, comprising:

[0032] An acquisition module is used to obtain real-time traffic data of a target area;

[0033] A model processing module, used to process the real-time traffic data through a pre-trained deep learning model to obtain a traffic prediction result, wherein the traffic prediction result represents the future traffic conditions of the target area;

[0034] A scheduling module is used to formulate and execute traffic scheduling strategies based on the traffic prediction results, wherein the traffic scheduling strategies include scheduling traffic nodes and / or displaying traffic guidance information, and the scheduling traffic nodes include adjusting the duration of traffic lights in the target area and / or regulating the roads in the target area.

[0035] An electronic device comprises: a processor, and a memory communicatively connected to the processor;

[0036] The memory stores computer-executable instructions;

[0037] The processor executes the computer-executable instructions stored in the memory to implement the traffic dispatch method as described in any one of the above items.

[0038] A computer-readable storage medium comprises: the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to implement the traffic dispatching method as described in any one of the above items when executed by a processor.

[0039] A computer program product comprises a computer program, wherein when the computer program is executed by a processor, the traffic dispatching method as described in any one of the above items is implemented.

[0040] The beneficial effects of the present invention are as follows: by pre-training the deep learning model with traffic data, the deep learning model can automatically learn traffic patterns and rules, and provide high-precision and real-time traffic prediction results based on real-time traffic data, thereby reducing the difficulty of traffic scheduling, so that effective scheduling can be carried out with the help of traffic prediction results, traffic congestion can be alleviated, and traffic conditions can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flow chart of a traffic dispatching method provided for an exemplary embodiment of the present invention;

[0042] Figure 2 A schematic diagram of a process for training a deep learning model provided by an exemplary embodiment of the present invention;

[0043] Figure 3 A schematic diagram of a process for processing real-time traffic data provided by an exemplary embodiment of the present invention;

[0044] Figure 4A schematic diagram of a flow chart for formulating a traffic dispatching strategy provided by an exemplary embodiment of the present invention;

[0045] Figure 5 A schematic diagram of an application scenario of a traffic dispatching method provided by an exemplary embodiment of the present invention;

[0046] Figure 6 A flow chart of a traffic dispatching scheme provided for an exemplary embodiment of the present invention;

[0047] Figure 7 A schematic diagram of the structure of a traffic dispatching device provided for an exemplary embodiment of the present invention;

[0048] Figure 8 A schematic structural diagram of an electronic device provided by an exemplary embodiment of the present invention.

[0049] Through the above drawings, clear embodiments of the present invention have been shown, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present invention in any way, but to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0050] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0051] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0052] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0053] The terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, product or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, product or device. In the absence of further restrictions, it is not excluded that there are other identical or equivalent elements in the process, method, product or device including the elements. For example, if the words first, second, etc. are used to indicate names, they do not indicate any particular order.

[0054] Figure 1 FIG. 1 is a flow chart of a traffic dispatching method provided by an exemplary embodiment of the present invention. Figure 1 As shown, the method may include:

[0055] Step S101, obtaining real-time traffic data of a target area.

[0056] The target area may be an area with complex traffic scenes such as urban roads. The sources of real-time traffic data may include vehicle positioning data, social media data, meteorological data, and data collected by preset sensors.

[0057] Specifically, data related to roads and vehicles can be collected in real time through tools such as road sensors and cameras deployed in the target area. Meteorological data, social media data and vehicle satellite positioning data can also be obtained through the Internet to comprehensively monitor the traffic environment.

[0058] For example, various sensors (such as vehicle detectors, geomagnetic sensors, etc.) can be deployed on major roads and intersections in the city to collect data such as traffic flow, vehicle speed, and lane occupancy in real time.

[0059] For example, high-definition cameras can be installed on major roads and key nodes to conduct video surveillance and identify information such as vehicle type, license plate number, and traffic accidents.

[0060] For example, by calling the system interfaces of taxi companies, bus companies and logistics companies, the real-time satellite positioning data of the vehicle can be obtained to understand the location, driving speed and driving route of the vehicle.

[0061] For example, traffic-related information (such as traffic accidents, road closures, emergencies, etc.) posted by users can be collected by calling the API (Application Programming Interface) of the social media platform.

[0062] Step S102: Process the real-time traffic data through a pre-trained deep learning model to obtain a traffic prediction result.

[0063] Among them, the traffic prediction results represent the future traffic conditions in the target area.

[0064] Specifically, a pre-trained deep learning model can be obtained. The deep learning model can be trained based on a large amount of traffic data generated by traffic scenarios similar to the target area. The real-time traffic data of the target area such as traffic flow, vehicle speed and meteorological information collected is input into the deep learning model. The model can output prediction results such as traffic flow, congestion and accident risk in the target area within a certain period of time in the future (such as the next hour).

[0065] In some possible implementations, before real-time traffic data is input into a deep learning model, the data can also be cleaned to remove noise and erroneous data (for example, abnormal data generated by sensor failures can be filtered out, duplicate and invalid satellite positioning data can be removed, etc.), and standardized data that matches the model input format can be generated.

[0066] Step S103: formulate and execute a traffic dispatch strategy based on the traffic prediction result.

[0067] The traffic dispatching strategy includes dispatching traffic nodes and / or displaying traffic guidance information. Dispatching traffic nodes includes adjusting the duration of traffic lights in the target area and / or regulating the roads in the target area.

[0068] Specifically, the traffic prediction results can be input into a trained neural network, which can determine the optimal adjustment strategy for traffic nodes in the target area based on information such as future traffic flow and congestion in the target area. For example, if the traffic node in the target area is a traffic light, the neural network can determine the optimal duration of different color signal lights of the traffic light in the future based on the input traffic flow. Controlling the operation of the traffic light based on the optimal duration can alleviate the congestion in the target area during this period. The neural network can be a BP (Back Propagation) neural network, which can include multiple hidden layers and a fully connected layer.

[0069] In some possible implementations, data visualization technology can also be used to display traffic forecast results in the form of charts, heat maps, etc., so as to manually formulate or optimize traffic scheduling strategies.

[0070] In some possible implementations, based on the traffic prediction results, a traffic dispatch strategy may be formulated through optimization algorithms such as genetic algorithms or particle swarm algorithms.

[0071] In some possible implementations, regulating roads in the target area may include closing roads, adjusting variable lanes, etc. For example, the target area may include one or more roads, and closing roads may be closing all or part of a road. For another example, the roads in the target area may include multiple lanes including variable lanes, and adjusting variable lanes may be changing the direction of the variable lanes allowed to pass during a certain period of time.

[0072] In the above embodiment, by acquiring the real-time traffic data of the target area, the real-time traffic data can be further processed by the pre-trained deep learning model, and the traffic prediction result can characterize the future traffic conditions of the target area, and a traffic dispatching strategy can be formulated and executed based on the obtained traffic prediction result. By pre-training the deep learning model with traffic data, the deep learning model can automatically learn traffic patterns and rules, and provide high-precision and real-time traffic prediction results based on real-time traffic data, reducing the difficulty of traffic dispatching, so that effective dispatching can be carried out with the help of traffic prediction results, traffic congestion can be alleviated, and traffic conditions can be improved.

[0073] In one embodiment, the traffic guidance information may include traffic prediction results and recommended routes determined based on the traffic prediction results.

[0074] In some possible implementations, displaying the traffic guidance information may include: displaying the traffic guidance information on a display screen within the target area.

[0075] For example, information display screens may be set up at major traffic nodes in a city to publish real-time traffic information, traffic forecast results, travel recommendations and other information through the display screens.

[0076] In some possible implementations, displaying the traffic guidance information may include: pushing the traffic guidance information to a terminal device used by the user.

[0077] For example, interactive tools such as mobile apps and mini-programs can be developed to push traffic information and navigation suggestions to users in real time. Traffic guidance information can also be integrated into the vehicle's navigation system to provide real-time route optimization and traffic warnings through the on-board system.

[0078] In one embodiment, Figure 2 As shown in the figure, the pre-trained deep learning model is obtained in the following way:

[0079] Step S201, obtaining historical traffic data.

[0080] The historical traffic data may be traffic data collected from the target area over a period of time in the past, or may be traffic data collected from other areas with traffic scenes similar to those of the target area.

[0081] Step S202: construct a training set and a validation set based on historical traffic data.

[0082] Step S203: training the initial model based on the data in the training set.

[0083] The initial model may be a deep learning model based on a bird's eye view fusion (bevfusion) architecture.

[0084] Step S204, adjusting the parameters of the trained initial model based on the data in the validation set to obtain a pre-trained deep learning model.

[0085] Specifically, the acquired historical traffic data can be preprocessed by data cleaning and other preprocessing operations, and a suitable deep learning model architecture can be selected according to the needs of traffic forecasting. The preprocessed data can be used to construct training sets and validation sets. The training set can be used for model training, and the validation set can be used for model evaluation and parameter adjustment.

[0086] For example, historical traffic data collected from the target area can be used as a validation set, and historical traffic data collected from other areas with traffic scenes similar to the target area can be used as a training set.

[0087] In this way, deep learning and reinforcement learning algorithms can be used to train a large multi-task model that can simultaneously predict traffic flow, congestion, and accident risks at multiple time steps. The model is continuously updated during training to adapt to changes in traffic conditions.

[0088] In one embodiment, the real-time traffic data includes point cloud features, image features, and indication information of traffic nodes in the target area, such as Figure 3 As shown, processing real-time traffic data through a pre-trained deep learning model may include:

[0089] Step S301 : Map the point cloud features and the image features to the three-dimensional space corresponding to the target area respectively, and determine the semantic information of the three-dimensional space according to the mapping results.

[0090] The point cloud features may be obtained based on a laser radar, and the image features may be extracted from images taken by cameras at different viewing angles.

[0091] Specifically, a BEV (bird's eye view) spatial model corresponding to the target area can be established. The spatial model can include several grids, and the point cloud features and image features can be mapped to the grids of the spatial model. The point cloud features and image features of each grid can be comprehensively analyzed to determine the semantic information of each grid.

[0092] Step S302: input the indication information and semantic information into a pre-trained deep learning model to obtain a traffic prediction result output by the deep learning model.

[0093] The indication information may include information such as the duration of traffic lights in the target area and road control rules.

[0094] Specifically, the deep learning model can perceive traffic information such as vehicle flow in the target area based on the input semantic information, and can combine the indication information such as the duration of traffic lights in the target area and variable lane rules to predict the changes in vehicle flow in the target area when traffic nodes such as traffic lights in the target area operate according to these indication information.

[0095] It should be noted that the indication information is the rules that the traffic nodes should follow when the traffic scheduling strategy is not implemented (such as the duration of traffic lights, variable lane directions, etc.).

[0096] In one embodiment, Figure 4 As shown in the figure, traffic dispatch strategies are formulated and implemented based on traffic forecast results, including:

[0097] Step S401, initializing the first generation of individuals based on the genetic algorithm.

[0098] Among them, the genes of the first generation of individuals represent the behavior of scheduling traffic nodes.

[0099] Specifically, a genetic algorithm can be used to generate an initial population of multiple individuals, each of which can represent the behavior of a possible traffic node, and the behavior of the traffic node can be encoded as the gene of the individual. For example, if the behavior of the traffic node is to control whether a road is closed, binary encoding can be used, where 0 indicates that the road is closed and 1 indicates that the road is open. If the behavior of the traffic node is to control the traffic light, real number encoding can be used, and the value of the gene can represent the green light duration of the traffic light (or other related parameters, such as cycle length, red-green ratio, etc.).

[0100] Step S402: determining a fitness function according to the traffic prediction result.

[0101] Specifically, at least one indicator may be selected from traffic flow, congestion index and accident risk according to the traffic prediction result; and a fitness function may be constructed according to the selected indicator.

[0102] For example, the traffic prediction results may include the traffic flow, congestion index and accident risk in the target area in the next ten minutes. The three indicators of traffic flow, congestion index and accident risk can be selected to construct the fitness function: F = A × traffic flow + B × congestion index + C × accident risk. Among them, F represents the fitness function value, A, B, and C are all weight coefficients, and the value of the weight coefficient can be dynamically adjusted according to the severity of different indicators in the traffic prediction results.

[0103] Step S403, iterating the first generation of individuals by changing the numerical value of the gene, and calculating the fitness value of each individual according to the fitness function.

[0104] Specifically, individuals can be selected from the initial population for reproduction by a roulette wheel method or the like, and the gene values ​​of the individuals can be changed by operations such as gene crossover and gene mutation to generate new individuals, and the fitness value of each individual can be calculated according to the fitness function. Gene crossover can include single-point crossover, two-point crossover, etc. Gene mutation can be to randomly change the gene value of an individual with a preset probability.

[0105] Step S404, selecting a target individual according to the fitness value until the termination condition of the iteration is reached.

[0106] Specifically, the termination condition of the iteration (such as reaching the maximum number of iterations, fitness value convergence, etc.) can be set in advance. Before the termination condition is reached, target individuals with higher or lower fitness values ​​are continuously selected during the iteration process and the gene values ​​are recorded.

[0107] Step S405, after the iteration is terminated, a traffic dispatch strategy is formulated according to the genes of the selected target individuals.

[0108] Specifically, after the iteration is terminated, the target individuals corresponding to each traffic node in the target area can be integrated, and the gene values ​​of these target individuals can be restored to the scheduling behavior of the traffic nodes, so as to formulate an overall traffic scheduling strategy based on the scheduling behavior.

[0109] In some possible implementations, a single traffic node may correspond to multiple target individuals. For example, there is a traffic light in the target area. Two target individuals corresponding to the traffic light are selected during the iteration process. Each target individual corresponds to a plan for the length of time the green light is on. These two plans can be combined with the scheduling behaviors of other traffic nodes, and the final applicable plan can be analyzed based on the combined traffic conditions.

[0110] To further demonstrate the beneficial effects of the present invention, Figure 5 and Figure 6 The specific application of the traffic dispatching method provided by the present invention is described.

[0111] Figure 5A schematic diagram of an application scenario of a traffic dispatch method provided by an exemplary embodiment of the present invention. According to the traffic dispatch method provided by the present invention, a traffic dispatch system for a target road scene can be designed, which can include data acquisition devices such as a plurality of sensors and cameras, and can also include devices for displaying information such as display screens, and servers running in the cloud.

[0112] like Figure 5 As shown, cameras, speed sensors, geomagnetic sensors (not shown in the figure) and other devices can be deployed on the roads within the target area (such as an intersection or a larger road environment). These devices can upload the collected data such as license plate numbers, vehicle speeds, number of vehicles, number of pedestrians and lane occupancy information to the cloud server through the network.

[0113] For traffic nodes such as traffic lights at intersections, the cloud server can not only obtain information such as the flashing duration of the traffic lights, but also send instructions to the traffic lights through the network to change the flashing rules of the traffic lights.

[0114] In addition, with the permission of the vehicle, the cloud server can also obtain the vehicle's precise satellite positioning information and other data through the network, and send detour suggestions and other traffic guidance information to the vehicle. For example, the user in the vehicle can run the preset APP (application) client through the vehicle system or the terminal device carried by the user, and the cloud server can run the APP server. The user can not only upload the vehicle's satellite positioning and driving route information through the client, but also upload the real-time congestion status of the area and other information. The server can aggregate multi-source data to perceive the traffic conditions in the area and generate optimization decisions (such as detour suggestions and other traffic guidance information or traffic light modification instructions, etc.). The server can also send the aggregated traffic conditions and traffic guidance information to the user side through the APP, or push the traffic conditions and traffic guidance information to the roadside display screen for display.

[0115] Figure 6 A flow chart of a traffic dispatching scheme provided for an exemplary embodiment of the present invention. The scheme can be applied to Figure 5 The road scene in the figure can also be applied to other scenes suitable for traffic scheduling.

[0116] like Figure 6As shown, according to the traffic dispatching scheme of this example, first, the multi-source traffic data of the target area can be collected through pre-set sensors such as cameras and calling third-party interfaces, and the data from different data sources can be integrated to build a training set and a validation set. The data in the training set can be used to train a large model based on deep learning, and the data in the validation set can be used to verify whether the training accuracy of the large model meets the requirements. If it does not meet the requirements, the model architecture and model parameters can be adjusted through the model selector, or the data in the training set can be updated until the trained large model meets the accuracy requirements and is used for traffic prediction. Then, the traffic dispatch decision is optimized based on the traffic prediction results, and the optimized decision is presented to the user through display screens, mobile applications, etc.

[0117] In the above example, the solution of the present invention utilizes multiple data sources to provide a more comprehensive and accurate analysis of traffic conditions. By introducing advanced technologies such as deep learning and reinforcement learning, the accuracy and reliability of traffic forecasts can be improved, so that traffic scheduling strategies can be adjusted in real time according to the prediction results, effectively alleviating traffic congestion and improving traffic efficiency and safety.

[0118] Figure 7 The following is a schematic diagram of a traffic dispatching device provided by an exemplary embodiment of the present invention. Figure 7 As shown, the traffic dispatching device 700 may include:

[0119] The acquisition module 701 is used to acquire real-time traffic data of a target area.

[0120] The model processing module 702 is used to process the real-time traffic data through a pre-trained deep learning model to obtain a traffic prediction result, and the traffic prediction result represents the future traffic conditions of the target area.

[0121] The scheduling module 703 is used to formulate and execute traffic scheduling strategies based on the traffic prediction results, wherein the traffic scheduling strategies include scheduling traffic nodes and / or displaying traffic guidance information, and the scheduling traffic nodes include adjusting the duration of traffic lights in the target area and / or regulating the roads in the target area.

[0122] In one embodiment, the scheduling module 703 may include:

[0123] The initialization unit can be used to initialize the first generation of individuals based on a genetic algorithm, and the genes of the first generation of individuals represent the behavior of the scheduling traffic node.

[0124] The function determination unit can be used to determine the fitness function according to the traffic prediction result.

[0125] The iteration unit can be used to iterate the first generation of individuals by changing the numerical value of the gene, and calculate the fitness value of each individual according to the fitness function.

[0126] The individual selection unit can be used to select a target individual according to the fitness value until the termination condition of the iteration is reached.

[0127] The strategy formulation unit can be used to formulate a traffic dispatch strategy based on the genes of the selected target individuals after the iteration is terminated.

[0128] In one embodiment, the function determination unit is further used to: select at least one indicator from traffic flow, congestion index and accident risk according to the traffic prediction result; and construct a fitness function according to the selected indicator.

[0129] In one embodiment, the model processing module 702 may include:

[0130] The mapping unit may be used to map the point cloud features and the image features to the three-dimensional space corresponding to the target area, respectively, and determine the semantic information of the three-dimensional space according to the mapping results.

[0131] The result acquisition unit can be used to input the indication information and the semantic information into a pre-trained deep learning model to obtain the traffic prediction result output by the deep learning model.

[0132] In one embodiment, the traffic dispatching device 700 may further include:

[0133] The historical acquisition module can be used to obtain historical traffic data.

[0134] The data construction module can be used to construct a training set and a validation set based on the historical traffic data.

[0135] The training module can be used to train the initial model based on the data in the training set, and the initial model is a deep learning model based on the bird's-eye view fusion architecture.

[0136] The parameter adjustment module can be used to adjust the parameters of the trained initial model based on the data in the verification set to obtain a pre-trained deep learning model.

[0137] In one embodiment, the scheduling module 703 is further used to: display the traffic guidance information on a display screen within the target area; and / or push the traffic guidance information to a terminal device used by a user.

[0138] The traffic dispatching device provided in this embodiment is used to execute the technical solution in any of the aforementioned method embodiments. Its implementation principle and technical effects are similar and will not be described in detail here.

[0139] It should be understood that the above device embodiments are only illustrative, and the device of the present invention can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0140] In addition, unless otherwise specified, each functional unit / module in each embodiment of the present invention may be integrated into one unit / module, or each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0141] Figure 8 FIG. 1 is a schematic diagram of a structure of an electronic device provided by an exemplary embodiment of the present invention. Figure 8 As shown, the electronic device 80 includes:

[0142] Processor 81, memory 82, and communication interface 83;

[0143] The memory 82 is used to store executable instructions of the processor 81; the executable instructions may be computer-executable instructions;

[0144] The processor 81 is configured to execute the technical solution in any of the aforementioned method embodiments by executing the executable instructions.

[0145] Optionally, the memory 82 can be independent or integrated with the processor 81.

[0146] Optionally, when the memory 82 is a device independent of the processor 81, the electronic device 80 may further include:

[0147] The bus 84 , the memory 82 and the communication interface 83 are connected to the processor 81 via the bus 84 and communicate with each other. The communication interface 83 is used to communicate with other devices.

[0148] Optionally, the communication interface 83 may be implemented by a transceiver. The communication interface is used to implement communication between the database access device and other devices (such as a client, a read-write library, and a read-only library). The memory may include a random access memory (RAM) and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0149] The bus 84 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0150] The above-mentioned 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, discrete hardware components.

[0151] The electronic device is used to execute the technical solution in any of the aforementioned method embodiments, and its implementation principle and technical effect are similar and will not be repeated here.

[0152] An embodiment of the present invention further provides a readable storage medium, which may be a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the technical solution provided by any of the aforementioned method embodiments is implemented.

[0153] An embodiment of the present invention further provides a computer program product, including a computer program, which is used to implement the technical solution provided by any of the aforementioned method embodiments when executed by a processor.

[0154] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0155] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not 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.

[0156] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. The description and examples are to be considered exemplary only, and the true scope and spirit of the present invention is indicated by the following claims.

[0157] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A traffic dispatching method, characterized in that: include: Obtain real-time traffic data for the target area; Processing the real-time traffic data through a pre-trained deep learning model to obtain a traffic prediction result, wherein the traffic prediction result represents the future traffic conditions of the target area; A traffic dispatching strategy is formulated and executed based on the traffic prediction results. The traffic dispatching strategy includes dispatching traffic nodes and / or displaying traffic guidance information. The dispatching traffic nodes includes adjusting the duration of traffic lights in the target area and / or regulating roads in the target area.

2. The traffic dispatching method according to claim 1, characterized in that: The formulating and executing a traffic dispatching strategy based on the traffic prediction result includes: Initialize the first generation of individuals based on the genetic algorithm, the genes of the first generation of individuals represent the behavior of the dispatching traffic node; Determining a fitness function according to the traffic prediction result; Iterating the first generation of individuals by changing the value of the gene, and calculating the fitness value of each individual according to the fitness function; Selecting a target individual according to the fitness value until an iteration termination condition is reached; After the iteration is terminated, a traffic dispatch strategy is formulated based on the genes of the selected target individuals.

3. The traffic dispatching method according to claim 2, characterized in that: The determining of the fitness function according to the traffic prediction result comprises: Selecting at least one indicator from traffic flow, congestion index and accident risk according to the traffic prediction result; Construct a fitness function based on the selected indicators.

4. The traffic dispatching method according to any one of claims 1 to 3, characterized in that: The real-time traffic data includes point cloud features, image features, and indication information of traffic nodes in the target area, and the processing of the real-time traffic data by using a pre-trained deep learning model includes: Mapping the point cloud features and the image features to the three-dimensional space corresponding to the target area respectively, and determining the semantic information of the three-dimensional space according to the mapping results; The indication information and the semantic information are input into a pre-trained deep learning model to obtain a traffic prediction result output by the deep learning model.

5. The traffic dispatching method according to any one of claims 1 to 3, characterized in that: The pre-trained deep learning model is obtained in the following way: Get historical traffic data; Constructing a training set and a validation set based on the historical traffic data; Training an initial model based on the data in the training set, wherein the initial model is a deep learning model based on a bird's eye view fusion architecture; The parameters of the trained initial model are adjusted based on the data in the validation set to obtain a pre-trained deep learning model.

6. The traffic dispatching method according to any one of claims 1 to 3, characterized in that: The displayed traffic guidance information includes: Displaying the traffic guidance information on a display screen within the target area; and / or, The traffic guidance information is pushed to the terminal device used by the user.

7. A traffic dispatching device, characterized in that: include: An acquisition module is used to obtain real-time traffic data of a target area; A model processing module, used to process the real-time traffic data through a pre-trained deep learning model to obtain a traffic prediction result, wherein the traffic prediction result represents the future traffic conditions of the target area; A scheduling module is used to formulate and execute traffic scheduling strategies based on the traffic prediction results, wherein the traffic scheduling strategies include scheduling traffic nodes and / or displaying traffic guidance information, and the scheduling traffic nodes include adjusting the duration of traffic lights in the target area and / or regulating the roads in the target area.

8. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

10. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 6 when being executed by a processor.

Images