A traffic control scene generation model construction and traffic control scene generation method

Abstract

The application discloses a traffic control scene generation model construction and traffic control scene generation method, acquires element entities used for composing a traffic control scene; determines a plurality of element keywords corresponding to each element entity; constructs a corresponding domain knowledge dataset according to the plurality of element keywords and extracts associated element keywords corresponding to each element keyword; randomly arranges and combines each element keyword and the associated element keywords corresponding thereto to obtain a plurality of combined scene data corresponding to each element keyword; inputs the plurality of combined scene data and the domain knowledge dataset into a preset scene generation model for training and obtains a traffic control scene generation model, and uses the model to determine whether a generated target traffic control scene is feasible. Through the application, the domain knowledge dataset is increased; the associated expression of the domain knowledge dataset and the combined scene data is realized; and the feasibility and accuracy of scene generation are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, specifically to a traffic control scenario generation model construction and a traffic control scenario generation method. Background Technology

[0002] Typically, intelligent traffic management and control centers such as "City Brain" and "Traffic Brain" handle fixed or pre-set traffic control scenarios, such as vehicle violations and accident handling. How to uniformly manage the scenarios and applications of the control center, and how to reasonably generate new scenarios, are the primary issues that cities need to address in building their "City Brain" and "Traffic Brain" platforms.

[0003] Existing methods for generating traffic control scenarios mainly include using adversarial neural networks to generate scenarios that approximate reality or relying on expert experience. However, the data used to train adversarial neural networks is limited, and relying on expert experience is highly subjective, resulting in low accuracy and feasibility of the generated traffic control scenarios and large errors compared to actual scenarios. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method for constructing a traffic control scenario generation model and generating traffic control scenarios, in order to solve the technical problems of low accuracy and feasibility of traffic control scenarios generated in the prior art, and large errors with actual scenarios.

[0005] The technical solution proposed in this invention is as follows:

[0006] The first aspect of this invention provides a method for constructing a traffic control scenario generation model. This method includes: acquiring element entities that constitute a traffic control scenario; determining multiple element keywords corresponding to each element entity; constructing a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords, and extracting associated element keywords corresponding to each element keyword from the domain knowledge dataset; randomly arranging and combining each element keyword and its associated element keywords to obtain multiple combined scenario data corresponding to each element keyword; and inputting the multiple combined scenario data containing element keywords and associated element keywords, along with the domain knowledge dataset, into a preset scenario generation model for training to obtain the traffic control scenario generation model.

[0007] Optionally, before constructing a corresponding domain knowledge dataset based on the element keywords and extracting associated element keywords corresponding to the element keywords in the domain knowledge dataset, the method further includes: acquiring historical traffic control business data and processing the historical traffic control business data; constructing a corresponding traffic control business set based on the processing result and associating the element entities with the traffic control business set; and constructing a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords and the associative traffic control business set.

[0008] A second aspect of the present invention provides a method for generating traffic control scenarios. The method includes: acquiring an element entity dataset for composing a traffic control scenario; inputting the element entity dataset into a traffic control scenario generation model constructed by the traffic control scenario generation model construction method as described in the first aspect and any one of the first aspects of the present invention, to obtain multiple target traffic control scenarios composed of the element entity dataset and a confidence level corresponding to each of the traffic control scenarios; and determining the multiple target traffic control scenarios whose confidence levels meet preset confidence requirements as feasible traffic control business scenarios.

[0009] Optionally, determining multiple target traffic control scenarios whose confidence levels meet preset confidence requirements as feasible traffic control business scenarios includes: comparing the element entity combination corresponding to each target traffic control scenario whose confidence levels meet preset confidence requirements with the domain knowledge dataset; filtering the multiple target traffic control scenarios whose confidence levels meet preset confidence requirements based on the comparison results, and taking the filtered target traffic control scenarios as feasible traffic control business scenarios.

[0010] Optionally, the method further includes: when multiple target traffic control scenarios whose confidence levels meet preset confidence requirements after screening are not feasible traffic control business scenarios, pruning is performed on the obtained element entity dataset; the element entity combination corresponding to multiple target traffic control scenarios generated based on the pruned element entity dataset is compared with the domain knowledge dataset, wherein the confidence level corresponding to the target traffic control scenario meets the preset confidence requirements; and the screening operation is repeated on the multiple target traffic control scenarios based on the comparison results.

[0011] Optionally, after repeatedly filtering the multiple target traffic control scenarios based on the comparison results, the method further includes: if there are still no feasible traffic control business scenarios among the multiple target traffic control scenarios generated after pruning, repeating the step of pruning the obtained element entity dataset to the step of repeatedly filtering the multiple target traffic control scenarios based on the comparison results, until the element entity dataset contains only one element entity.

[0012] Optionally, the method further includes: when there is still no feasible traffic control business scenario among the plurality of target traffic control scenarios, sending the element entity dataset to the user terminal; when the user terminal determines that any target traffic control scenario composed of the element entity dataset is the feasible traffic control business scenario, adding the element entity combination corresponding to the target traffic control scenario to the domain knowledge dataset.

[0013] A third aspect of this invention provides a traffic control scenario generation model construction device, comprising: a first acquisition module for acquiring element entities used to compose a traffic control scenario; a determination module for determining multiple element keywords corresponding to each element entity; an extraction module for constructing a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords and extracting associated element keywords corresponding to each element keyword from the domain knowledge dataset; a combination module for randomly arranging and combining each element keyword and its associated element keywords to obtain multiple combined scenario data corresponding to each element keyword; and a training module for inputting multiple combined scenario data containing element keywords and associated element keywords and the domain knowledge dataset into a preset scenario generation model for training to obtain a traffic control scenario generation model.

[0014] A fourth aspect of the present invention provides a traffic control scenario generation device, comprising: a second acquisition module for acquiring an element entity dataset for composing a traffic control scenario; an input module for inputting the element entity dataset into a traffic control scenario generation model constructed by the traffic control scenario generation model construction method as described in the first aspect and any one of the first aspects of the present invention, to obtain multiple target traffic control scenarios composed of the element entity dataset and a confidence level corresponding to each of the traffic control scenarios; and a determination module for determining multiple target traffic control scenarios whose confidence levels meet preset confidence requirements as feasible traffic control business scenarios.

[0015] A fifth aspect of the present invention provides a computer-readable storage medium storing computer instructions for causing the computer to execute the traffic control scenario generation model construction method as described in the first aspect and any one of the first aspects of the present invention, or the traffic control scenario generation method as described in the second aspect and any one of the second aspects of the present invention.

[0016] A sixth aspect of the present invention provides an electronic device, including: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the traffic control scenario generation model construction method as described in the first aspect and any one of the first aspects of the present invention, or the traffic control scenario generation method as described in the second aspect and any one of the second aspects of the present invention.

[0017] The technical solution provided by this invention has the following effects:

[0018] The traffic control scenario generation model construction method provided in this embodiment of the invention utilizes multiple element keywords corresponding to element entities to construct a domain knowledge dataset corresponding to each element keyword, thereby increasing the domain knowledge dataset; the traffic control scenario generation model is obtained by training with the domain knowledge dataset and combined scenario data, realizing the associated expression of the domain knowledge dataset and combined scenario data.

[0019] The traffic control scenario generation method provided in this embodiment of the invention uses a trained traffic control scenario generation model to obtain the confidence level corresponding to each traffic control scenario, and obtains the corresponding feasible traffic control business scenario based on the confidence level comparison, thereby improving the feasibility and accuracy of scenario generation. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a traffic control scenario generation model construction method according to an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of an autoencoder provided according to an embodiment of the present invention;

[0023] Figure 3This is a schematic diagram illustrating the association effect between element entities and traffic control business sets according to an embodiment of the present invention;

[0024] Figure 4 This is a flowchart of a traffic control scenario generation method according to an embodiment of the present invention;

[0025] Figure 5 This is a structural block diagram of a traffic control scenario generation model building device according to an embodiment of the present invention;

[0026] Figure 6 This is a structural block diagram of a traffic control scenario generation device according to an embodiment of the present invention;

[0027] Figure 7 This is a schematic diagram of the structure of a computer-readable storage medium provided according to an embodiment of the present invention;

[0028] Figure 8 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] This invention provides a method for constructing a traffic control scenario generation model, such as... Figure 1 As shown, the method includes the following steps:

[0031] Step S101: Obtain the element entities used to compose the traffic control scenario. Specifically, element entities can be combined to describe the traffic system, and may include, but are not limited to, personnel, vehicles, time, space, weather, facilities, rules, and other element entities. Define the relationships between elements based on their hierarchical or subordinate relationships; defining these relationships constitutes the traffic control scenario diagram.

[0032] Step S102: Determine multiple element keywords corresponding to each element entity. Specifically, multiple element keywords can completely describe the corresponding element entity. For example, when the element entity is a vehicle, the multiple element keywords corresponding to the vehicle can be: car, bus, truck, dangerous goods transport vehicle, autonomous vehicle, non-motorized vehicle, motorcycle, bus, taxi, ride-hailing vehicle; when the element entity is a person, the multiple element keywords corresponding to the person can be: driver, pedestrian, passenger; the multiple element keywords corresponding to the element entity rule can be: illegal, violation; wherein, the element keywords include, but are not limited to, existing data and types supervised by relevant departments such as traffic, traffic police, and municipalities, as well as data types that are obtainable but not yet obtained, or may be obtained, such as Internet text data, etc.

[0033] The corresponding traffic control scenarios can be obtained by combining the various element entities. For example, when there are 7 element entities and each element entity has N keywords, the total number of corresponding traffic control scenarios is N. 7 indivual.

[0034] Step S103: Construct a domain knowledge dataset corresponding to each type of element keyword based on the multiple element keywords, and extract the associated element keywords corresponding to each type of element keyword from the domain knowledge dataset. Specifically, taking Internet text data mining as an example, element keywords are used to search for web page links, the text content of the top n web pages ranked by relevance is saved, and words and word frequencies related to the content of each element entity in the text are extracted to form a domain knowledge dataset.

[0035] In one embodiment, a web search is performed using "dump truck" as the keyword element. The obtained text is statistically analyzed, and high-frequency but non-essential words such as "road," "traffic," "truck," and "safety" are deleted to obtain the corresponding domain knowledge dataset. The following related keyword elements can be extracted from the domain knowledge dataset: {overload, speeding / low speed, overloading, license plate obscured / covered, running red lights, violating traffic restrictions, overloading, illegal modification, illegal passenger transport, illegal parking, fatigued driving, safety around schools, reckless driving, spillage, vibration and noise, road hazards (such as obstructed visibility, road surface damage, dangerous bridges, high-risk road sections, icy roads), fire safety, high-risk enterprises, driving in the wrong direction, parking in the wrong direction, unlicensed driving, transporting without protective measures, not driving according to regulations, high travel demand, turning off GPS / falsifying data, breaking down in the middle of the road, blind spot, outstanding violations, mixed passenger and freight transport, staged accidents, overdue inspection}, totaling 31 types. The scenario is true when dump trucks and the above 31 situations occur together.

[0036] Step S104: Randomly arrange and combine each element keyword and its corresponding associated element keywords to obtain multiple combined scenario data corresponding to each element keyword. Specifically, each element keyword and its corresponding associated element keywords are randomly arranged and combined to obtain multiple combined scenario data.

[0037] For example, randomly arranging and combining the keyword "dump truck" with its corresponding related keywords can yield multiple different combination scenarios such as "dump truck overloaded" and "dump truck running a red light".

[0038] Step S105: Input multiple combined scene data containing element keywords and related element keywords, along with the domain knowledge dataset, into a preset scene generation model for training to obtain a traffic control scene generation model. Specifically, the scenes composed of the domain knowledge dataset are associated with actual scenes. Therefore, after inputting multiple combined scene data into the preset scene generation model and training it using the domain knowledge dataset, a corresponding traffic control scene generation model can be obtained. In this way, the traffic control scene obtained by this traffic control scene generation model is compared with the actual scene, improving the accuracy of the obtained traffic control scene.

[0039] The specific training methods can be traditional machine learning methods, deep learning, neural networks, etc., and are not specifically limited in this invention, as long as they meet the requirements. The preset scene generation model can be a word embedding model that can convert word vectors into computable structured vectors, such as the Word2Vec model or the BERT model. These are not specifically limited in this invention, as long as they meet the requirements.

[0040] In one embodiment, an autoencoder is used as an example. Based on the Word2Vec model, word vectors are reduced to a two-dimensional space and input into a... Figure 2 The autoencoder shown is designed to combine encoder-decoder reconstruction error with the reliability of low-dimensional space representation of the scene, and the parameters of the network are learned using a domain knowledge dataset.

[0041] Network training error includes, but is not limited to, the following:

[0042] 1) Reconstruction error of word vector representation:

[0043] loss1=(y out -y in ) 2

[0044] In the formula, y out y represents the output vector of the autoencoder; in This represents the input vector of the autoencoder.

[0045] 2) Scene determination error: In the last layer of network training, the low-dimensional space of the text is normalized to the range [0, 1] using the sigmoid function, i.e.: h represents the low-dimensional space; o is the normalized numerical value. The correctness of the scene determination is a binary classification problem: J∈{0,1};

[0046] Therefore, the scene determination error can be expressed as:

[0047]

[0048] In the formula, n represents the nth determination; j n This indicates the result of the nth determination, which is either 0 or 1; o n This represents the normalized value corresponding to the nth determination.

[0049] Specifically, network training error can be obtained by directly summing or weighted summing the reconstruction and scene determination errors.

[0050] The traffic control scenario generation model construction method provided in this embodiment of the invention utilizes multiple element keywords corresponding to element entities to construct a domain knowledge dataset corresponding to each element keyword, thereby increasing the domain knowledge dataset; the traffic control scenario generation model is obtained by training with the domain knowledge dataset and combined scenario data, realizing the associated expression of the domain knowledge dataset and combined scenario data.

[0051] As an optional implementation of this invention, before step S103, the method further includes: acquiring historical traffic control business data and processing the historical traffic control business data; constructing a corresponding traffic control business set based on the processing result and associating the element entity with the traffic control business set; and constructing a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords and the traffic control business set after association processing.

[0052] Specifically, firstly, historical traffic control business data includes, but is not limited to, existing traffic control business data from relevant departments such as transportation, traffic police, and municipalities. This traffic control business data will be integrated and divided into directions such as command, control, enforcement, supervision, and service. Each direction will be further divided into different scenarios and businesses, and corresponding traffic control business sets will be constructed.

[0053] Secondly, the element entities are associated with each scenario business in the traffic control business cluster, realizing the combined use of basic static data and measured data.

[0054] In one embodiment, the traffic control service set = {accident handling, driver management, operation monitoring, key road section analysis...}; the element entities are personnel, vehicles, time, space, weather, facilities, and rules; such as Figure 3 The image shown is a schematic diagram illustrating the effect of associating an element entity with the traffic control business set.

[0055] This invention also provides a method for generating traffic control scenarios, such as... Figure 4 As shown, the method includes the following steps:

[0056] Step S201: Obtain the dataset of element entities used to compose the traffic control scenario.

[0057] Step S202: Input the element entity dataset into the traffic control scenario generation model constructed by the traffic control scenario generation model construction method described in this embodiment of the invention, to obtain multiple target traffic control scenarios composed of the element entity dataset and the confidence level corresponding to each traffic control scenario. Specifically, element entities are used to compose traffic control scenarios; therefore, inputting the element entity dataset into the traffic control scenario generation model can yield multiple target traffic control scenarios.

[0058] Secondly, such as Figure 2 As shown, after inputting into the traffic control scenario generation model, the element entity dataset is mapped to the corresponding word vectors. These vectors are then encoded through a fully connected layer and decoded through a fully connected layer of the corresponding number to reconstruct the word vector combinations. Finally, the output value in the low-dimensional space is used as the confidence score for each traffic control scenario. The output value format is: Rule(["condition 1", "condition 2", "condition 3", ...], val), where val represents the confidence score; "condition 1", "condition 2", and "condition 3" represent the element combinations corresponding to the traffic control scenarios.

[0059] Step S203: Multiple target traffic control scenarios that meet the preset confidence level requirements are identified as feasible traffic control business scenarios. Specifically, the feasibility of the generated multiple target traffic control scenarios is determined using the number of eigenvalues ​​(i.e., confidence level) in the low-dimensional space. The target traffic control scenario is defined in the interval [0, 1]. The closer the value of the target traffic control scenario is to 1, the more feasible it is considered. In other words, multiple target traffic control scenarios that meet the preset confidence level requirements are identified as feasible traffic control business scenarios. If multiple target traffic control scenarios are determined to be infeasible, the process ends directly.

[0060] The traffic control scenario generation method provided in this embodiment of the invention uses a trained traffic control scenario generation model to obtain the confidence level corresponding to each traffic control scenario, and obtains the corresponding feasible traffic control business scenario based on the confidence level comparison, thereby improving the feasibility and accuracy of scenario generation.

[0061] As an optional implementation of this invention, after step S203, the method further includes: comparing the element entity combination corresponding to each target traffic control scenario whose confidence level meets the preset confidence level requirements with the domain knowledge dataset; filtering multiple target traffic control scenarios whose confidence level meets the preset confidence level requirements based on the comparison results, and taking the filtered target traffic control scenarios as feasible traffic control business scenarios.

[0062] Specifically, the domain knowledge dataset is searched to see if it contains a combination of element entities corresponding to the target traffic control scenario. If it does, the corresponding target traffic control scenario is determined to be a feasible traffic control business scenario. By comparison, the feasible traffic control business scenarios obtained in step S203 can be filtered, which improves the accuracy of the determination of feasible traffic control business scenarios, further improves the feasibility of the generated traffic control business scenarios, and reduces the error with the actual scenario.

[0063] If, after screening, multiple target traffic control scenarios whose confidence levels meet the preset confidence requirements are not feasible traffic control business scenarios, the obtained element entity dataset is pruned. The element entity combinations corresponding to the multiple target traffic control scenarios generated from the pruned element entity dataset are compared with the domain knowledge dataset, wherein the confidence levels corresponding to the target traffic control scenarios meet the preset confidence requirements. The screening operation for the multiple target traffic control scenarios is repeated based on the comparison results.

[0064] Specifically, when the domain knowledge dataset does not contain the element entity combination corresponding to the target traffic control scenario, the obtained element entity dataset is pruned, that is, one or more element entities are randomly deleted, and the domain knowledge dataset is searched again to check whether it contains the element entity combination corresponding to the target traffic control scenario. If it does, the corresponding target traffic control scenario is determined to be a feasible traffic control business scenario.

[0065] If, after pruning, there is still no feasible traffic control business scenario among the multiple target traffic control scenarios generated, repeat the step of pruning the obtained element entity dataset to the step of repeatedly filtering the multiple target traffic control scenarios based on the comparison results, until the element entity dataset contains only one element entity.

[0066] Specifically, when searching and checking again in the domain knowledge dataset, if the domain knowledge dataset still does not contain the combination of element entities corresponding to the target traffic control scenario, the pruning process for the element entity dataset is repeated and the search and checking in the domain knowledge dataset continues until there is only one element entity in the element entity dataset.

[0067] After iterative processing, if none of the element entity combinations corresponding to the target traffic control scenario are found in the domain knowledge dataset, meaning that none of the multiple target traffic control scenarios composed of the obtained element entity datasets are feasible traffic control business scenarios, then the method further includes: sending the element entity dataset to the user terminal; when the user terminal determines that any target traffic control scenario composed of the element entity dataset is the feasible traffic control business scenario, adding the element entity combination corresponding to the target traffic control scenario to the domain knowledge dataset.

[0068] Specifically, the human-computer interface is used to ask the operator whether there are feasible traffic control business scenarios among the multiple target traffic control scenarios composed of the element entity dataset. If the operator believes that there are any one or more feasible traffic control business scenarios among the multiple target traffic control scenarios, the element entity combination corresponding to the target traffic control scenario is added to the constructed domain knowledge dataset for subsequent judgment and use.

[0069] This invention also provides a device for constructing a traffic control scenario generation model, such as... Figure 5 As shown, the device includes:

[0070] The first acquisition module 501 is used to acquire element entities that constitute the traffic control scenario; for details, please refer to the relevant description of step S101 in the above method embodiment.

[0071] The determination module 502 is used to determine multiple element keywords corresponding to each element entity; for details, please refer to the relevant description of step S102 in the above method embodiment.

[0072] Extraction module 503 is used to construct a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords, and to extract the associated element keywords corresponding to each element keyword in the domain knowledge dataset; for details, please refer to the relevant description of step S103 in the above method embodiment.

[0073] The combination module 504 is used to randomly arrange and combine each of the element keywords and its corresponding associated element keywords to obtain multiple combination scene data corresponding to each of the element keywords; for details, please refer to the relevant description of step S104 in the above method embodiment.

[0074] The training module 505 is used to input multiple combined scene data containing element keywords and related element keywords and the domain knowledge dataset into a preset scene generation model for training and to obtain a traffic control scene generation model; for details, please refer to the relevant description of step S105 in the above method embodiment.

[0075] The traffic control scenario generation model construction device provided in this embodiment of the invention utilizes multiple element keywords corresponding to element entities to construct a domain knowledge dataset corresponding to each element keyword, thereby increasing the domain knowledge dataset; the traffic control scenario generation model is obtained by training with the domain knowledge dataset and combined scenario data, realizing the associated expression of the domain knowledge dataset and combined scenario data.

[0076] As an optional embodiment of the present invention, the apparatus further includes: a first processing module, configured to acquire historical traffic control business data and process the historical traffic control business data; a second processing module, configured to construct a corresponding traffic control business set based on the processing result and associate the element entity with the traffic control business set; and a construction module, configured to construct a domain knowledge dataset corresponding to each element keyword based on the plurality of element keywords and the associated traffic control business set.

[0077] For a detailed description of the function of the traffic control scenario generation model construction device provided in this embodiment of the invention, please refer to the description of the traffic control scenario generation model construction method in the above embodiments.

[0078] This invention also provides a traffic control scenario generation device, such as... Figure 6 As shown, the device includes:

[0079] The second acquisition module 601 is used to acquire the element entity dataset used to compose the traffic control scenario; for details, please refer to the relevant description of step S201 in the above method embodiment.

[0080] Input module 602 is used to input the element entity dataset into the traffic control scenario generation model constructed by the traffic control scenario generation model construction method as described in the embodiment of the present invention, to obtain multiple target traffic control scenarios composed of the element entity dataset and the confidence level corresponding to each traffic control scenario; for details, please refer to the relevant description of step S202 in the above method embodiment.

[0081] The determination module 603 is used to determine multiple target traffic control scenarios that meet the preset confidence requirements as feasible traffic control business scenarios; for details, please refer to the relevant description of step S203 in the above method embodiment.

[0082] The traffic control scenario generation device provided in this embodiment of the invention uses a trained traffic control scenario generation model to obtain the confidence level corresponding to each traffic control scenario, and obtains the corresponding feasible traffic control business scenario based on the confidence level comparison, thereby improving the feasibility and accuracy of scenario generation.

[0083] As an optional implementation of the present invention, the device further includes: a first comparison module, used to compare the element entity combination corresponding to each target traffic control scenario whose confidence level meets the preset confidence level requirement with the domain knowledge dataset; and a filtering module, used to filter multiple target traffic control scenarios whose confidence level meets the preset confidence level requirement according to the comparison results, and to select the target traffic control scenarios obtained after filtering as feasible traffic control business scenarios.

[0084] As an optional implementation of this invention, the determination module includes: a first processing submodule, used to prune the obtained element entity dataset when multiple target traffic control scenarios whose confidence levels meet preset confidence requirements after screening are not feasible traffic control business scenarios; a first comparison submodule, used to compare the element entity combinations corresponding to multiple target traffic control scenarios generated from the pruned element entity dataset with the domain knowledge dataset, wherein the confidence levels corresponding to the target traffic control scenarios meet preset confidence requirements; and a first determination submodule, used to repeatedly perform screening operations on the multiple target traffic control scenarios based on the comparison results.

[0085] As an optional implementation of the present invention, the device further includes: a repeating module, used to repeat the step of pruning the obtained element entity dataset to the step of repeating the filtering operation of the multiple target traffic control scenarios based on the comparison results when there is still no feasible traffic control business scenario among the multiple target traffic control scenarios generated after pruning, until the element entity dataset contains only one element entity.

[0086] As an optional implementation of this invention, the device further includes: a second determination module, configured to send the element entity dataset to the user terminal when there is still no feasible traffic control business scenario among the plurality of target traffic control scenarios; and an addition module, configured to add the element entity combination corresponding to the target traffic control scenario to the domain knowledge dataset when the user terminal determines that any target traffic control scenario composed of the element entity dataset is the feasible traffic control business scenario.

[0087] For a detailed description of the functions of the traffic control scenario generation device provided in this embodiment, please refer to the description of the traffic control scenario generation method in the above embodiments.

[0088] This invention also provides a storage medium, such as... Figure 7 As shown, a computer program 701 is stored on it. When executed by a processor, this program implements the steps of the traffic control scenario generation model construction method or traffic control scenario generation method in the above embodiments. The storage medium also stores audio and video stream data, feature frame data, interactive request signaling, encrypted data, and preset data size, etc. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium may also include combinations of the above types of memory.

[0089] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0090] This invention also provides an electronic device, such as... Figure 8 As shown, the electronic device may include a processor 81 and a memory 82, wherein the processor 81 and the memory 82 may be connected via a bus or other means. Figure 8 Taking the example of a connection between China and Israel via a bus.

[0091] Processor 81 can be a central processing unit (CPU). Processor 81 can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0092] The memory 82, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the corresponding program instructions / modules in the embodiments of the present invention. The processor 81 executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory 82, thereby implementing the traffic control scenario generation model construction method or traffic control scenario generation method in the above method embodiments.

[0093] The memory 82 may include a program storage area and a data storage area. The program storage area may store applications required for operating the device and at least one function; the data storage area may store data created by the processor 81, etc. Furthermore, the memory 82 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory 82 may optionally include memory remotely located relative to the processor 81, and these remote memories may be connected to the processor 81 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0094] The one or more modules are stored in the memory 82, and when executed by the processor 81, they perform the following: Figure 1 -4 shows the traffic control scenario generation model construction method or traffic control scenario generation method.

[0095] For specific details regarding the aforementioned electronic devices, please refer to the relevant documentation. Figures 1 to 4 The relevant descriptions and effects in the illustrated embodiments are for understanding purposes only and will not be repeated here.

[0096] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for constructing a traffic control scenario generation model, characterized in that, Includes the following steps: Obtain the element entities used to compose the traffic control scenario; Determine multiple element keywords corresponding to each of the aforementioned element entities; Construct a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords, and extract the associated element keywords corresponding to each element keyword from the domain knowledge dataset; Randomly arrange and combine each of the aforementioned element keywords and its corresponding associated element keywords to obtain multiple combined scenario data corresponding to each of the aforementioned element keywords; Multiple combination scene data containing element keywords and related element keywords, along with the domain knowledge dataset, are input into a preset scene generation model for training to obtain a traffic control scene generation model. The preset scene generation model reduces word vectors to a two-dimensional space, inputs them into an autoencoder, and combines the encoder-decoder reconstruction error with the reliability design of the low-dimensional space representing the scene. The parameters of the model are learned using the domain knowledge dataset. The model training error includes the reconstruction error of the word vector representation and the scene judgment error. The scene judgment error is determined based on the scene judgment result, which is a binary classification result used to represent whether the scene judgment is correct or not. The domain knowledge dataset is obtained through the following steps: acquiring historical traffic control business data and processing the historical traffic control business data; Based on the processing results, construct the corresponding traffic control business set and associate the element entities with the traffic control business set; Based on the multiple element keywords and the traffic control business set after association processing, construct a domain knowledge dataset corresponding to each element keyword.

2. A method for generating traffic control scenarios, characterized in that, Includes the following steps: Obtain the dataset of element entities used to compose the traffic control scenario; The element entity dataset is input into the traffic control scenario generation model constructed by the traffic control scenario generation model construction method as described in claim 1, to obtain multiple traffic control scenarios composed of the element entity dataset and the confidence level corresponding to each traffic control scenario. Multiple target traffic control scenarios that meet the pre-set confidence level requirements are identified as feasible traffic control business scenarios.

3. The method according to claim 2, characterized in that, The step of determining multiple target traffic control scenarios that meet the preset confidence level requirements as feasible traffic control business scenarios includes: The element entity combination corresponding to each target traffic control scenario whose confidence level meets the preset confidence level requirements is compared with the domain knowledge dataset; Based on the comparison results, multiple target traffic control scenarios that meet the preset confidence level requirements are screened, and the screened target traffic control scenarios are taken as feasible traffic control business scenarios.

4. The method according to claim 3, characterized in that, The method further includes: If, after screening, multiple target traffic control scenarios that meet the preset confidence requirements are not feasible traffic control business scenarios, the obtained element entity dataset is pruned. The element entity combinations corresponding to multiple target traffic control scenarios generated from the pruned element entity dataset are compared with the domain knowledge dataset, wherein the confidence level of the target traffic control scenario meets the preset confidence level requirements. Based on the comparison results, the screening operation for the multiple target traffic control scenarios is repeated.

5. The method according to claim 4, characterized in that, After repeatedly filtering the multiple target traffic control scenarios based on the comparison results, the method further includes: If there is still no feasible traffic control business scenario among the multiple target traffic control scenarios generated after pruning, repeat the step of pruning the obtained element entity dataset to the step of repeatedly filtering the multiple target traffic control scenarios based on the comparison results, until the element entity dataset contains only one element entity. If no feasible traffic control business scenario exists among the multiple target traffic control scenarios, the element entity dataset is sent to the user terminal. When a user terminal determines that any target traffic control scenario composed of the element entity dataset is the feasible traffic control business scenario, the element entity combination corresponding to the target traffic control scenario is added to the domain knowledge dataset.

6. A traffic control scenario generation model construction device, characterized in that, include: The first acquisition module is used to acquire the element entities that make up the traffic control scenario; The determination module is used to determine multiple element keywords corresponding to each of the element entities; The extraction module is used to construct a domain knowledge dataset corresponding to each element keyword based on the multiple element keywords, and to extract the associated element keywords corresponding to each element keyword from the domain knowledge dataset. The combination module is used to randomly arrange and combine each of the element keywords and its corresponding associated element keywords to obtain multiple combination scene data corresponding to each of the element keywords. The training module is used to input multiple combined scene data containing element keywords and related element keywords and the domain knowledge dataset into a preset scene generation model for training and to obtain a traffic control scene generation model. The preset scene generation model reduces word vectors to a two-dimensional space, inputs them into an autoencoder, and combines the design of encoder-decoder reconstruction error and low-dimensional space to represent the reliability of the scene. The parameters of the model are learned using the domain knowledge dataset. The model training error includes reconstruction error of word vector representation and scene judgment error. The scene judgment error is determined according to the scene judgment result. The scene judgment result is a binary classification result used to represent whether the scene judgment is correct or not. The domain knowledge dataset is obtained through the following steps: acquiring historical traffic control business data and processing the historical traffic control business data; Based on the processing results, construct the corresponding traffic control business set and associate the element entities with the traffic control business set; Based on the multiple element keywords and the traffic control business set after association processing, construct a domain knowledge dataset corresponding to each element keyword.

7. A traffic control scenario generation device, characterized in that, include: The second acquisition module is used to acquire the dataset of element entities that make up the traffic control scenario; The input module is used to input the element entity dataset into the traffic control scenario generation model constructed by the traffic control scenario generation model construction method as described in claim 1, so as to obtain multiple traffic control scenarios composed of the element entity dataset and the confidence level corresponding to each traffic control scenario. The determination module is used to determine multiple target traffic control scenarios that meet the preset confidence requirements as feasible traffic control business scenarios.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the traffic control scenario generation model construction method as described in claim 1, or the traffic control scenario generation method as described in any one of claims 2-5.

9. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the traffic control scenario generation model construction method as described in claim 1, or the traffic control scenario generation method as described in any one of claims 2-5.

Images