Traffic control scheme generation method and related device, equipment and storage medium

Abstract

The application discloses a signal control scheme generation method and related devices, equipment and storage medium, wherein the signal control scheme generation method comprises: obtaining a signal control planning text of a target object for a target intersection; wherein the signal control planning text is a natural language planning of traffic signal switching change of the target object; constructing a large model instruction based on the signal control planning text; wherein the large model instruction is used to instruct a large language model to analyze the signal control planning text into structured data about traffic phases; obtaining output data of the large language model responding to the large model instruction as target data; generating a signal control scheme of the target intersection based on the target data; wherein the signal control scheme comprises control parameters of each traffic phase in a cycle length. The above scheme can improve the generation efficiency of the signal control scheme and as much as possible improve the operation convenience in the signal control scheme generation process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traffic control, in particular to a traffic control scheme generation method and related device, equipment and storage medium. BACKGROUND

[0002] Generally, traffic flow is usually allocated by traffic signals at traffic intersections to improve the safety, orderliness and smoothness of traffic travel.

[0003] Currently, traffic control schemes need to be manually configured for traffic intersections, such as manually inputting various parameters of each traffic phase of a traffic intersection. However, due to the large number of traffic intersections in real scenarios and the complexity of some intersections, manual configuration is extremely inefficient and cumbersome. Therefore, how to improve the generation efficiency of traffic control schemes and as much as possible to improve the operation convenience in the traffic control scheme generation process has become a problem to be solved. SUMMARY

[0004] The technical problem solved by the present application is to provide a traffic control scheme generation method and related device, equipment and storage medium, which can improve the generation efficiency of traffic control schemes and as much as possible to improve the operation convenience in the traffic control scheme generation process.

[0005] To solve the above technical problem, the first aspect of the present application provides a traffic control scheme generation method, comprising: obtaining a traffic control planning text of a target object for a target intersection; wherein the traffic control planning text is a natural language planning of the jump change of the traffic signal by the target object; constructing a large model instruction based on the traffic control planning text; wherein the large model instruction is used to instruct a large language model to analyze the traffic control planning text into structured data about traffic phases; obtaining output data of the large language model responding to the large model instruction as target data; generating a traffic control scheme of the target intersection based on the target data; wherein the traffic control scheme includes control parameters of each traffic phase in a cycle length.

[0006] To solve the above technical problem, the second aspect of the present application provides a traffic control scheme generation device, comprising: a planning acquisition module, an instruction construction module, a data acquisition module and a scheme generation module, the planning acquisition module is used to obtain a traffic control planning text of a target object for a target intersection; wherein the traffic control planning text is a natural language planning of the jump change of the traffic signal by the target object; the instruction construction module is used to construct a large model instruction based on the traffic control planning text; wherein the large model instruction is used to instruct a large language model to analyze the traffic control planning text into structured data about traffic phases; the data acquisition module is used to obtain output data of the large language model responding to the large model instruction as target data; the scheme generation module is used to generate a traffic control scheme of the target intersection based on the target data; wherein the traffic control scheme includes control parameters of each traffic phase in a cycle length.

[0007] To solve the above technical problems, the third aspect of the present application provides an electronic device, comprising at least a memory and a processor coupled with each other, the memory at least stores program instructions, and the processor is configured to execute the program instructions to implement the signal control scheme generation method in the first aspect.

[0008] To solve the above technical problems, the fourth aspect of the present application provides a computer readable storage medium, which stores program instructions capable of being executed by a processor, and the program instructions are used to implement the signal control scheme generation method in the first aspect.

[0009] The above scheme, the signal control planning text of the target object to the target intersection is obtained, and the signal control planning text is the jump change of the traffic signal planned by the target object in natural language. Then, based on the signal control planning text, a large model instruction is constructed, and the large model instruction is used to instruct the large language model to analyze the signal control planning text into structured data about traffic phases, so as to obtain the output data of the large language model responding to the large model instruction as the target data. Then, based on the target data, the signal control scheme of the target intersection is generated, and the signal control scheme includes the control parameters of each traffic phase in the cycle time. Therefore, on the one hand, the structured data is generated by analyzing the signal control planning text by the large language model to generate the signal control scheme, which can improve the generation efficiency of the signal control scheme compared with manually inputting each parameter of each traffic phase. On the other hand, the target object can describe the signal control planning of the target intersection in natural language during the signal control planning process, which can avoid performing a large number of manual operations and can improve the operation convenience as much as possible during the signal control scheme generation process. Therefore, the generation efficiency of the signal control scheme can be improved, and the operation convenience during the signal control scheme generation process can be improved as much as possible. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 is a flowchart of an embodiment of the signal control scheme generation method of the present application;

[0011] Figure 2a is a schematic diagram of an embodiment of the traffic phase of the target intersection of the present application;

[0012] Figure 2b is a schematic diagram of an embodiment of the scheme assembly of the present application;

[0013] Figure 2c is a light color schematic diagram of an embodiment of the signal control scheme of the present application;

[0014] Figure 3 is a frame schematic diagram of an embodiment of the signal control scheme generation device of the present application;

[0015] Figure 4 is a frame schematic diagram of an embodiment of the electronic device of the present application;

[0016] Figure 5 FIG. 1 is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0017] The scheme of the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0018] In the following description, specific details are set forth in order to provide a thorough understanding of the present application. However, persons having ordinary skill in the art will appreciate that the present application can be practiced without the specific details.

[0019] The terms "system" and "network" are often used interchangeably herein. The term "and / or", merely describes an associated relationship between associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the following three cases: A exists alone, A and B exist together, and B exists alone. In addition, the segment " / " herein generally represents an "or" relationship between the associated objects. In addition, "multiple" herein means two or more than two.

[0020] Please refer to Figure 1 , Figure 1 FIG. 1 is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of the present application.

[0021] Step S11: Obtain the traffic control planning text of the target object for the target intersection.

[0022] In the embodiments of the present disclosure, the traffic control planning text is the jump change of the traffic signal planned by the target object in natural language. For example, the traffic control planning text can be the jump change of the periodic traffic signal control described by the target object in natural language. As a possible implementation example, the traffic control planning text can be obtained by the target object through character input, such as the target object inputting the traffic control planning text through pinyin input, handwriting input, etc. Alternatively, as another possible implementation example, the traffic control planning text can be obtained by the target object through voice input, such as the target object describing how to plan the jump change of the traffic signal through voice input, and then recognizing the voice data input by the target object, so as to obtain the traffic control planning text. Of course, the above examples are only several possible ways of obtaining the traffic control planning text in actual application, and other ways of obtaining the traffic control planning text are not limited herein, and will not be described one by one.

[0023] In an implementation scenario, the signal control planning text can describe how the traffic phases of the target intersection change with respect to the traffic signal. It should be noted that different traffic phases correspond to different traffic flows. For example, taking the target intersection as a crossroad, the traffic phase "South Left Turn Phase" corresponds to the traffic flow of motor vehicles turning left at the crossroad at the south entrance of the crossroad, the traffic phase "North Straight Phase" corresponds to the traffic flow of motor vehicles straight at the crossroad at the north entrance of the crossroad, and the traffic phase "East Pedestrian" corresponds to the traffic flow of pedestrians at the east entrance of the crossroad going south (or north). The above examples are only a few possible examples of traffic phases, and traffic phases will not be listed one by one here. For specific details of traffic phases in the field of signal control, please refer to the technical details of traffic phases in the field of signal control, which will not be described here.

[0024] In an implementation scenario, as described above, the signal control planning text can describe how the traffic phases of the target intersection change with respect to the traffic signal. The traffic phases in the signal control planning text can be referred to by phase names, or the traffic phases in the signal control planning text can also be referred to by phase numbers, which are not limited here. Taking the target intersection as a crossroad, different traffic phases can be referred to by phase names such as "North Right Turn", "North Left Turn", "North Straight", "South Right Turn", "South Left Turn", "South Straight", "East Right Turn", "East Left Turn", "East Straight", "West Right Turn", "West Left Turn", "West Straight", "North U-turn", "South U-turn", "East U-turn", "West U-turn", "North Pedestrian", "South Pedestrian", "East Pedestrian", "West Pedestrian"; or different traffic phases can also be referred to by phase numbers such as "φ1", "φ2", "φ3", "φ4", "φ5", "φ6", "φ7", "φ8", "φ9", "φ10", "φ11", "φ12", "φ13", "φ14", "φ15", "φ16". Of course, the above examples are only a few possible examples of referring to traffic phases in the signal control planning text, and other referring methods are not limited here, and will not be listed one by one.

[0025] In an implementation scenario, the interaction interface between the target object and the large language model can display a schematic diagram of the target intersection, and the schematic diagram can mark each traffic phase with at least one of a phase name and a phase number, so that the target object can refer to the schematic diagram to output the signal control planning text in natural language. Please refer to Figure 2a , Figure 2a is a schematic diagram of an embodiment of the traffic phases of the target intersection of the present application. As Figure 2aAs shown, taking the target intersection as an example, the intersection is a crossroad, and the schematic diagram can be marked with the phase number "φ1" of the traffic phase with the phase name "North Left Turn", the phase number "φ2" of the traffic phase with the phase name "North Straight", the phase number "φ3" of the traffic phase with the phase name "East Left Turn", the phase number "φ4" of the traffic phase with the phase name "East Straight", the phase number "φ5" of the traffic phase with the phase name "South Left Turn", the phase number "φ6" of the traffic phase with the phase name "South Straight", the phase number "φ7" of the traffic phase with the phase name "West Left Turn", the phase number "φ8" of the traffic phase with the phase name "West Straight", the phase number "φ10" of the traffic phase with the phase name "East Right Turn", the phase number "φ11" of the traffic phase with the phase name "South Right Turn", the phase number "φ12" of the traffic phase with the phase name "West Right Turn", the phase number "φ14" of the traffic phase with the phase name "North U-Turn", the phase number "φ15" of the traffic phase with the phase name "South U-Turn", and the phase number "φ16" of the traffic phase with the phase name "West U-Turn". When the pedestrian passage of the target intersection is provided with a pedestrian signal lamp for secondary crossing and a pedestrian crossing safety island, the pedestrian passage of the target intersection supports secondary crossing, that is, pedestrians can complete crossing in two segments. For example, the "North Pedestrian" "φA" can further include secondary crossing with the phase number "φE" and secondary crossing with the phase number "φF", the "East Pedestrian" "φB" can further include secondary crossing with the phase number "φG" and secondary crossing with the phase number "φH", the "South Pedestrian" "φC" can further include secondary crossing with the phase number "φI" and secondary crossing with the phase number "φJ", and the "West Pedestrian" "φD" can further include secondary crossing with the phase number "φK" and secondary crossing with the phase number "φL". Of course, Figure 2a The above-mentioned is only an example of indicating traffic phases by phase names and phase numbers when taking a crossroad as an example, and other possible cases are not limited herein. As a specific example, the signal control planning text can include but is not limited to the following content: "cycle 110 seconds, first ring South Left Turn 21 seconds, North Straight 44 seconds, second ring North Left Turn 39 seconds, South Straight 26 seconds. Then East-West Left Turn green light 17 seconds, East-West Straight green light 22 seconds. All phases 3 seconds of flashing green. South Right Turn light control, follow East Left Turn. East Right Turn from 88 seconds to 21 seconds. Add pedestrian follow-up phase" and the like, and the specific content of the signal control planning text is not limited herein.

[0026] Step S12: constructing a large model instruction based on the signal control planning text.

[0027] In the embodiments of the present disclosure, the large model instruction is used to instruct the large language model to parse the signal control planning text into structured data about traffic phases. For example, the structured data can include but is not limited to data formats such as XML, Json, and the like, and the data format of the structured data is not limited herein.

[0028] In one implementation scenario, as one possible implementation example, the large model instruction can be directly constructed according to the signal control planning text. For example, before the large language model is applied to the parsing task of the planning text, a sample planning text and its corresponding sample structured data about traffic signals can be obtained, so that the sample instruction of the large language model can be constructed based on the sample planning text, and the sample instruction is used to instruct the large language model to parse the sample planning text into predicted structured data about traffic phases. Then, based on the difference between the predicted structured data and the sample structured data, the parameters of the large language model can be fine-tuned. In this way, the large language model can be forced to learn how to accurately parse the structured data about traffic phases in the parameter fine-tuning process through a small amount of fine-tuning data (i.e., the sample planning text and its corresponding sample structured data). On this basis, when the large language model is applied to the parsing task of the planning text, the large model instruction can be directly constructed according to the signal control planning text, so that the large language model responds to the large model instruction to output the structured data about traffic phases.

[0029] In another implementation scenario, as another possible implementation example, in addition to the foregoing implementation, the prompt instruction template of the large language model can also be constructed based on prompt word engineering. The prompt instruction template at least includes several reference signal control plans of different plan structures, the processing methods of different traffic phases, the data output content and data output format of the large language model, and the large model instruction is constructed based on the prompt instruction template and the signal control planning text, so as to be used for the large language model to parse the signal control planning text by referring to the prompt instruction template. Through the prompt word engineering, the prompt instruction template is constructed first, and then the large model instruction is constructed based on the prompt instruction template and the signal control planning text, so that the large language model can learn the related domain knowledge before parsing, thereby assisting the subsequent parsing task.

[0030] In a specific implementation scenario, different plan structures can include but are not limited to: a ring plan structure, a stage plan structure, and the like, without limitation to the specific types of different plan structures. It should be noted that the ring plan structure is implemented based on ring boundaries, and common control forms are single ring or double ring, one or more rings can be included in one signal period, and one or more phases can be included in one ring. The corresponding traffic release phases need to be isolated and belong to different boundaries, such as the aforementioned signal control planning text “first ring south left turn 21 seconds, north straight 44 seconds, second ring north left turn 39 seconds, south straight 26 seconds”, which belongs to the ring plan structure. Correspondingly, the stage plan structure is controlled according to the phase stage protocol, and its control plan includes specific information of each stage, which focuses on dividing the signal control process into different stages, each stage corresponds to a specific phase combination and release rule, and the traffic signal is controlled by sequentially executing these stages. For example, the aforementioned signal control planning text “then east-west left turn green light 17 seconds, east-west straight green light 22 seconds” belongs to the stage plan structure. In addition to the above, a hybrid nested plan structure of ring and stage can also be used, such as first ring south left turn 21 seconds, north straight 44 seconds, second ring north left turn 39 seconds, south straight and south right turn 26 seconds. In addition, the several reference signal control planning can be common planning text examples of different plan structures, so that the large language model can understand the knowledge before performing the parsing task.

[0031] In a specific implementation scenario, the processing mode of different traffic phases can include but is not limited to: the processing mode of left turn phases, right turn phases, pedestrian phases, and the like, how to record the following phases (the specific meaning of which can be referred to in the subsequent description, which will not be described here) and the like, without limitation to the specific contents of the processing mode of different traffic phases, and will not be described one by one.

[0032] In one specific implementation scenario, as one possible implementation example, the data output of the large language model can include, but is not limited to, a first structure (which can contain phase parameters of the main phase, and record the scheme structure and phase order formed by the main phase), a second structure (which can contain phase parameters of the main phase, the following phase of the main phase, the independent phase independent of the main phase, the pedestrian phase, and the like), a third structure (which can contain the cycle length), and the like, without limitation. Of course, each structure in the data output of the large language model also stores data in a structured form, such as in the form of key-value (i.e., key-value pair), without limitation. It should be noted that the main phase, the following phase, and the independent phase are only relative concepts, and do not mean that a specific traffic phase (such as "south left turn", "north straight", and the like) must be the main phase or the following phase. Specifically, the main phase is contained in each stage or ring of the scheme structure and represents the phase of the stage or ring. For example, the aforementioned signal control planning text "first ring south left turn 21 seconds, north straight 44 seconds, second ring north left turn 39 seconds, south straight 26 seconds" belongs to a ring scheme structure, and "south left turn", "north straight", "north left turn", and "south straight" are the main phases in this example (and are the main phases contained in the ring). The aforementioned signal control planning text "then east-west left turn green light 17 seconds, east-west straight green light 22 seconds" belongs to a stage scheme structure, and "east left turn", "west left turn", "east straight", and "west straight" are the main phases in this example (and are the main phases contained in the stage). Unlike the main phase, the following phase of the main phase refers to the traffic phase attached to the main phase. The following phase and the main phase it follows are usually in the same stage. For example, the aforementioned signal control planning text "south right turn light control, following east left turn" means that "south right turn" is a following phase and its main phase is "east left turn", and the aforementioned signal control planning text "add a pedestrian following phase" means that the pedestrian phase is also a following phase, but does not specify the main phase that the pedestrian phase follows. In addition, the independent phase, as its name implies, usually does not have a dependent relationship with the main phase. For example, the aforementioned signal control planning text "east right turn from 88 seconds to 21 seconds" directly specifies the start and end time of the traffic phase "east right turn", and does not depend on any other traffic phase. Of course, the above examples are only a few possible examples of the main phase, the following phase, and the independent phase in actual application, and other possible scenarios are not limited herein and will not be listed one by one. In addition, in order to facilitate understanding of the keys that may be involved in the signal control field, please refer to Table 1, which is a schematic table of key (key) of phase attribute one embodiment. Of course, other keys (keys) may also be provided in addition to the keys (keys) listed in Table 1, which will not be listed one by one.

[0033] Table 1: Illustrative table of keys for phase attributes, one embodiment

[0034]

[0035] In one specific implementation scenario, the data output format of the large language model can include but is not limited to structured data formats such as XML, Json, etc. as described above, and the specific type of output format is not limited here.

[0036] Step S13: Obtain the output data of the large language model responding to the large model instruction as target data.

[0037] In one implementation scenario, when the traffic phase in the signal control planning text involves a main phase, the large language model can obtain the scheme structure describing the main phase in the signal control planning text as the target structure when responding to the large model instruction to perform the parsing task, and the scheme structure can include at least one of a ring type and a stage type. Based on this, the structure body created by the large language model for the target structure can be obtained as the first structure body, and the target data includes the first structure body. The above-mentioned manner, when the traffic phase in the signal control planning text involves a main phase, the large language model sequentially parses the scheme structure describing the main phase, creates the corresponding structure body, and updates according to the phase parameter, which can as much as possible clearly reflect each scheme structure involved in the target data in the form of a structure body. The main phase.

[0038] In one specific implementation scenario, still taking the above-mentioned signal control planning text as an example, as described above, "the first ring south left turn 21 seconds, north straight 44 seconds, the second ring north left turn 39 seconds, south straight 26 seconds" belongs to the scheme structure of the ring type, and "then east-west left turn green light 17 seconds, east-west straight green light 22 seconds" belongs to the scheme structure of the stage type, so the signal control planning text contains two scheme structures of the stage type and the ring type. The scheme structure of the stage type and the scheme structure of the ring type can be selected as the target structure, respectively. Of course, the above-mentioned example is only one possible example in the actual application process, and other possible cases are not limited here, and will not be exemplified one by one.

[0039] In one specific implementation scenario, the first structure can record the phase parameters in a list structure (e.g., a double-layer list, i.e., a list of lists). In addition, in order to mark whether the scheme structure to which the first structure belongs is ring type or stage type, the first structure created by the large language model itself can also contain a structure mark marking whether the scheme structure to which it belongs is ring type or stage type. Exemplarily, the structure mark can be "stageStyle" (indicating stage type) or "ringStyle" (indicating ring type).

[0040] In one specific implementation scenario, after (or at the same time, or before) creating the structure for the target structure, the large language model can also parse the main phase in the signal control planning text to the phase parameters described by the target structure. It should be noted that in this scenario, the phase parameters can include but are not limited to the phase duration of the main phase, etc. In addition, in the case where the phase parameters include the phase duration, the phase duration can specifically refer to the complete duration including green light, yellow light, red light, etc., or can refer to only the green light duration, which is not limited here. Of course, in order to distinguish between the two cases, different keys can be used, such as in the case where the phase duration represents the complete duration described above, the key "split" can be used to mark, and in the case where the phase duration represents the green light duration described above, the key "greenTime" can be used to mark. Of course, the above examples are only a few possible examples of the phase parameters of the main phase, and other possible cases are not limited here, nor are they exemplified one by one.

[0041] In one specific implementation scenario, in order to facilitate understanding, still taking the aforementioned signal control planning text as an example, as described above, "first ring south left turn 21 seconds, north straight 44 seconds, second ring north left turn 39 seconds, south straight 26 seconds" belongs to the ring type scheme structure, and "then east-west left turn green light 17 seconds, east-west straight green light 22 seconds" belongs to the stage type scheme structure. For the former, the following first structure "result1" can be created and updated:

[0042] "ringStyle": [ [

[0044] { "south left turn": { "split": 21}},

[0045] { "north straight": { "split": 44}}

[0046] ], [

[0048] { "north left turn": { "split": 39}},

[0049] {"southStraight": { "split": 26}} ] ]

[0052] The first structure represents the phase duration (complete duration) of the main phase "south left turn" in the ring scheme structure as 21 seconds, the phase duration (complete duration) of the main phase "north straight" as 44 seconds, which need to be controlled in sequence. In addition, the phase duration (complete duration) of the main phase "north left turn" is 39 seconds, and the phase duration (complete duration) of the main phase "south straight" is 26 seconds, which also need to be controlled in sequence. For the latter, the following first structure "result1" can be created and updated:

[0053] "stageStyle": [ [

[0055] {"eastLeft": { "greenTime": 17}},

[0056] {"westLeft": { "greenTime": 17}}

[0057] ], [

[0059] {"eastStraight": { "greenTime": 22}},

[0060] {"westStraight": { "greenTime": 22}} ] ]

[0063] The first structure represents the phase duration (green duration) of the main phase "east left turn" in the stage scheme structure as 17 seconds, the phase duration (green duration) of the main phase "west left turn" as 17 seconds, which need to be controlled in sequence. In addition, the phase duration (green duration) of the main phase "east straight" is 22 seconds, and the phase duration (green duration) of the main phase "west straight" is 22 seconds, which also need to be controlled in sequence. Of course, the above example is only two possible examples of the first structure, and other possible cases are not limited here, and will not be exemplified one by one.

[0064] In one implementation scenario, the target data can include a second structure created by the large language model for recording phase attribute of the record, and when the large language model performs the parsing task in response to the large model instruction, at least one of the following can be performed: in response to the large language model detecting that the traffic phase in the signal control planning text involves a main phase, based on the phase parameters of the main phase other than the phase duration parsed by the large language model in the signal control planning text, updating the phase attribute of the main phase in the second structure; in response to the large language model detecting that the traffic phase in the signal control planning text involves a following phase of the main phase, based on the main phase followed by the following phase parsed by the large language model in the signal control planning text, updating the phase attribute of the following phase in the second structure; in response to the large language model detecting that the traffic phase in the signal control planning text involves an independent phase independent of the main phase, based on the start and end time of the independent phase parsed by the large language model in the signal control planning text, updating the phase attribute of the independent phase in the second structure; in response to the large language model detecting that the traffic phase in the signal control planning text involves a pedestrian phase, based on the phase parameters of the pedestrian phase parsed by the large language model in the signal control planning text, updating the phase attribute of the pedestrian phase in the second structure; wherein the phase parameters of the pedestrian phase include the main phase followed by the pedestrian phase in the signal control planning text, or when the main phase followed by the pedestrian phase is not specified in the signal control planning text, the phase parameters of the pedestrian phase are a default placeholder. It should be noted that the above examples are only possible examples of updating the phase attribute of the corresponding phase in the second structure when the large language model detects that the traffic phase in the signal control planning text involves a main phase, a following phase of the main phase, an independent phase independent of the main phase, and a pedestrian phase, respectively, and other possible scenarios are not limited here, nor are they listed one by one. Of course, in a real scenario, the above several possible examples can exist simultaneously in the same signal control planning text, or only some of the possible examples (such as one of the possible examples, or two of the possible examples, or three of the possible examples) can exist in the same signal control planning text, which is not limited here.

[0065] In one specific implementation scenario, as one possible example, in the case where the large language model detects that the traffic phase in the signal control planning text involves a main phase, the phase attribute of the main phase in the second structure can be updated based on the phase parameters other than the phase duration of the main phase parsed by the large language model in the signal control planning text. It should be noted that in a real scenario, in addition to the phase duration of the main phase in the signal control planning text, there can be no other phase parameters, and then the second structure can be updated according to the default values of the keys such as greenFlash and phaseOrder in the signal control field as shown in Table 1. In order to facilitate understanding, still taking the aforementioned signal control planning text as an example, as mentioned earlier, the main phase can include "South Left Turn", "North Straight", "North Left Turn", "South Straight", "East Left Turn", "West Left Turn", "East Straight", "West Straight", since these eight have no phase parameters other than the green flash duration (specified as 3 seconds in the aforementioned signal control planning text) and the phase duration in the signal control planning text, the default values of the keys such as phaseOrder can be used to update the second structure, such as the following key-value pairs in the second structure "result2":

[0066] "result2":[

[0067] {"South Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0068] {"North Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0069] {"North Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0070] {"South Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0071] {"East Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0072] {"West Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0073] {"East Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0074] {"West Straight": {"phaseOrder": 1, "greenFlash": 3}} ]

[0076] In one specific implementation scenario, as one possible example, in the case where the large language model detects that the traffic phase in the traffic control planning text involves a following phase of a main phase, the phase attribute of the following phase in the second structure can be updated based on the main phase that the following phase follows, which is parsed by the large language model in the traffic control planning text. As shown in the aforementioned Table 1, the key “refPhase” can be used to mark the main phase that the following phase follows. In addition, as described above for the main phase with respect to the phase duration other than the phase parameters, for the following phase, the second structure can also be updated according to the default values of the keys such as greenFlash, phaseOrder, etc. in the traffic control field as shown in Table 1. For ease of understanding, still taking the aforementioned traffic control planning text as an example, as previously described, the main phase that the following phase “South Right Turn” follows is “East Left Turn”, then the following key-value pair can be updated in the second structure “result2”:

[0077] "result2":[

[0078] {"South Right Turn": {"phaseOrder": 1, "greenFlash": 3, "refPhase": "East Left Turn-1"}} ]

[0080] It should be noted that the number followed by the hyphen “-” in the above example represents the stage that the following phase needs to follow the main phase that it follows. In addition, if the target object does not declare this in the traffic control planning text, the first stage can be followed by default.

[0081] In one specific implementation scenario, as one possible example, in the case where the large language model detects that the traffic phase in the traffic control planning text involves an independent phase independent of the main phase, the phase attribute of the independent phase in the second structure can be updated based on the start and end time of the independent phase parsed by the large language model in the traffic control planning text. In addition, as described above for the main phase with respect to the phase duration other than the phase parameters, for the independent phase, the second structure can also be updated according to the default values of the keys such as greenFlash, phaseOrder, etc. in the traffic control field as shown in Table 1. For ease of understanding, still taking the aforementioned traffic control planning text as an example, as previously described, from the aforementioned traffic control planning text “East Right Turn from 88 seconds to 21 seconds”, the start and end time of the independent phase “East Right Turn” can be parsed as 88 seconds and 21 seconds respectively, then the following key-value pair can be updated in the second structure “result2”:

[0082] "result2":[

[0083] {"East Right Turn": {"phaseOrder": 1, "greenFlash": 3, "startTime": 88, "endTime": 21}} ]

[0085] In one specific implementation scenario, as one possible example, in the case where the large language model detects that the traffic phase in the signal control planning text involves a pedestrian phase, the phase attribute of the pedestrian phase in the second structure can be updated based on the phase parameter of the pedestrian phase parsed by the large language model in the signal control planning text. It should be noted that the phase parameter of the pedestrian phase can include the main phase followed by the pedestrian phase specified in the signal control planning text, or, in the case where the main phase followed by the pedestrian phase is not specified in the signal control planning text, the large language model can directly record the phase parameter of the pedestrian phase as a default placeholder (such as default, etc.) to prevent the large language model from making mistakes in determining the main phase followed by the pedestrian phase based on autonomous reasoning. As previously described, the key "refPhase" can be used to mark the main phase followed by the pedestrian phase (or marked as a default placeholder). In addition, as previously described for the phase parameters other than the phase duration of the main phase, for the pedestrian phase, the second structure can also be updated according to the default values of the keys (such as greenFlash, phaseOrder, etc.) in the signal control field as shown in Table 1. For ease of understanding, still taking the aforementioned signal control planning text as an example, as previously described, "adding a pedestrian following phase" from the aforementioned signal control planning text can parse that the pedestrian phase does not specify the main phase followed, so the following key-value pairs can be updated in the second structure "result2":

[0086] "result2":[

[0087] {"North Pedestrian": {"phaseOrder": 1, "refPhase": "default"}},

[0088] {"South Pedestrian": {"phaseOrder": 1, "refPhase": "default"}},

[0089] {"East Pedestrian": {"phaseOrder": 1, "refPhase": "default"}},

[0090] {"West Pedestrian": {"phaseOrder": 1, "refPhase": "default"}} ]

[0092] It should be noted that the final second structure of the aforementioned signal control plan text can be obtained by merging the second structures updated according to the aforementioned different situations. Exemplarily, the final second structure can be represented as:

[0093] "result2":[

[0094] {"South Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0095] {"North Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0096] {"North Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0097] {"South Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0098] {"East Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0099] {"West Left Turn": {"phaseOrder": 1, "greenFlash": 3}},

[0100] {"East Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0101] {"West Straight": {"phaseOrder": 1, "greenFlash": 3}},

[0102] {"South Right Turn": {"phaseOrder": 1, "greenFlash": 3, "refPhase": "East Left Turn-1"}},

[0103] {"East Right Turn": {"phaseOrder": 1, "greenFlash": 3, "startTime": 88, "endTime": 21}},

[0104] {"North Pedestrian": {"phaseOrder": 1, "refPhase": "default"}},

[0105] {"South Pedestrian": {"phaseOrder": 1, "refPhase": "default"}},

[0106] {"East Traveler":{"phaseOrder":1 , "refPhase":"default"}},

[0107] {"Westbound Traveler":{"phaseOrder":1 , "refPhase":"default"}} ]

[0109] In a specific implementation scenario, to reduce the computational load on large language models and lower their size requirements, making them more suitable for smaller, less powerful models with fewer parameters, the design of prompt words can focus on leveraging the semantic understanding capabilities of the large model and minimizing complex reasoning and calculations. For example, for common traffic control measures, prompt words can be used to directly record these as key-value pairs, rather than performing multi-step complex reasoning and calculations to generate the result. For instance, when describing "Northbound straight ahead 15 seconds late" in traffic control planning text, it can be directly recorded in the "Northbound straight ahead" structure as: {"lateStart":15}, without requiring the large model to add 15 seconds to the original green light start time to recalculate the actual green light start time.

[0110] In one implementation scenario, the target data may contain a third structure created by the large language model to record the cycle duration. When the large language model performs parsing on the traffic control planning text, it can also update the third structure based on the cycle duration parsed from the text if it detects that a cycle duration has been parsed. Alternatively, it can update the third structure with a target identifier if it detects that a cycle duration has not been parsed from the text. The target identifier represents the unspecified cycle duration in the traffic control planning text. For example, the target identifier can be null, etc. No specific limitations are imposed on the target identifier here, and further examples will not be provided. For ease of understanding, using the aforementioned traffic control planning text as an example, the large language model can parse the specified cycle duration of 110 seconds from "cycle 110 seconds" in the aforementioned traffic control planning text. Therefore, the third structure "result3" can be updated with the following key-value pairs:

[0111] "result3":110

[0112] Of course, the above examples are only one possible example of the third structure in the actual application process, and other possible cases are not limited here, nor are they listed one by one. For example, if the large language model does not parse the cycle length in the signal control planning text, the third structure can be directly updated with the target identifier null, such as "result3": null.

[0113] Step S14: Based on the target data, generate a signal control scheme for the target intersection.

[0114] In the embodiments of the present disclosure, the signal control scheme includes the control parameters of each traffic phase in the cycle length. It should be noted that for any traffic phase, the control parameters of the traffic phase in the cycle length can include: the control light color (such as red, green, yellow) of the traffic phase at each time in the cycle length. In addition, the process steps of generating a signal control scheme according to the structured data about the traffic phase can be programmed into a program product (such as a plug-in) using Python or other programming languages, so that the structured target data is read by the program product and the signal control scheme is generated.

[0115] In one implementation scenario, taking the example of the traffic phase involving the main phase, the follow-up phase of the main phase, and the independent phase independent of the main phase in the signal control planning text, the control parameters of each main phase in the cycle length can be determined based on the scheme structure of the main phase in the target data. It should be noted that, as described earlier, the scheme structure can include at least one of ring type and stage type. On this basis, the control parameters of the target phase in the cycle length can be determined based on the control parameters of the main phase in the cycle length and the main phase followed by the target phase in the target data, and the control parameters of the independent phase in the cycle length can be determined based on the start and end time of the independent phase in the target data. It should be noted that the target phase at least includes the follow-up phase. In addition, as one possible example, the target phase can also include the pedestrian phase. It should be noted that the control parameters at least include the start and end time. In addition, as one possible example, the control parameters can also specifically include green light start and end time, yellow light start and end time, red light start and end time, etc. It should also be noted that the above examples are only one possible example of generating a signal control scheme based on the main phase first and then the follow-up phase and the independent phase in the case of the traffic phase involving the main phase, the follow-up phase and the independent phase in the signal control planning text. In real scenarios, the traffic phase in the signal control planning text can only involve the main phase, or only involve the main phase and its follow-up phase, or only involve the main phase and the independent phase, etc. In these different cases, the signal control scheme can be generated by referring to the above-mentioned manner respectively, and here it will not be described one by one.

[0116] In one specific implementation scenario, when processing the main phases, the start and end time of the main phases in the cycle length can be obtained by sequentially connecting each plan structure in the target data based on the phase length of the main phases in the plan structure. In addition, for each main phase itself (e.g., the main phase "south left turn" in the aforementioned signal control planning text), the phase attributes (e.g., green flash length) of the main phase can be extracted from the second structure of the target data, and the default phase attributes (e.g., yellow light length) can be extracted, to further determine the green start and end time, green flash start and end time, yellow light start and end time of the main phase in the cycle length in combination with the start and end time of the main phase in the cycle length, and the remaining length is the red light part. For ease of understanding, still taking the aforementioned signal control planning text as an example, please refer to Figure 2b , Figure 2b is a schematic diagram of an embodiment of the present application. As Figure 2bAs shown, since the "ring" scheme structure in the target data, it can be determined that the start and end time of the main phase "South Left" and "North Straight" in the cycle length is from 0 to 65 seconds, and the start and end time of the former is from 0 to 21 seconds, and the start and end time of the latter is from 21 to 65 seconds, and similarly, the start and end time of the second ring main phase "North Left" and "South Straight" in the cycle length is also from 0 to 65 seconds, the difference is that the start and end time of the main phase "North Left" in the second ring is from 0 to 39 seconds, and the start and end time of the main phase "South Straight" is from 39 to 65 seconds; further, for the "stage" scheme structure in the target data, the first stage main phase "East Left" and "West Left" can be sequentially connected to the end time "65 seconds" of the first ring and the second ring, that is, the two main phases in the first stage start from "65 seconds", and the green light time of the two main phases in the first stage is 17 seconds, plus the default yellow light time "3 seconds", that is, the phase time of the two main phases in the first stage is 20 seconds, that is, the two main phases in the first stage end at "85 seconds"; similarly, the second stage main phase "East Straight" and "West Straight" are sequentially connected to the end time "85 seconds" of the two main phases in the first stage, that is, the two main phases in the second stage start from "85 seconds", and the green light time of the two main phases in the second stage is "22 seconds", plus the default yellow light time "3 seconds", that is, the phase time of the two main phases in the second stage is "25 seconds", that is, the two main phases in the second stage end at "110 seconds", that is, a cycle length "110 seconds" in the above example, at this time, the specific value of the cycle length does not need to be updated by using the total time of the sequential connection of the phase time of the main phase, on the contrary, if the total time is not consistent with the specific value of the cycle length in the target data (or the cycle length is not specified in the target data), the specific value of the cycle length can be updated based on the total time. It should be noted that the above example is only one possible example of scheme assembly for main phases in actual application, and other possible cases are not limited. For example, there can be multiple stage structures in the target data, and if the same main phase exists in multiple stage structures, the start and end time of the main phase in the cycle length can be determined by the phase time defined by the main phase in different stage structures.It should be noted that the connectivity of the phase stage can specifically require that two stages of the same phase have an intersection (or strict connection) in time, and that the same phase is a protected phase (such as a straight phase, that is, ensuring that other phases conflicting with it are all red light states) or a permissive phase (such as a left turn phase, that is, if a green light is on when a left turn green light is not configured, the opposite traffic flow is also observed). In addition, as previously described, after determining the time of all main phases, the specific value of the cycle length can be updated, such as can be equal to the maximum value of the end time of all main phases.

[0117] In a specific implementation scenario, when processing the target phase (as previously described, at least can include the following phase, in addition to the pedestrian phase), the phase parameter of the main phase followed by the target phase in the target data and unrelated to any of the late start and early termination can be obtained as the target parameter, and the control parameter of the target phase in the cycle length can be determined based on the target parameter and the phase parameter of the target phase in the target data and related to any of the late start and early termination. Specifically, in the second structure of the target data, all phases with "refPhase" not empty and not main phase, that is, the following phase, can be obtained. For example, in the example of the aforementioned "result2", the traffic phase "South Right Turn" has "refPhase" not empty and not main phase, so it is a following phase. On this basis, the main phase followed by the following phase can be determined based on the phase name, phase number and other phase attributes in the "refPhase" attribute of the main phase, and the start time and end time of the following phase are the same as the main phase followed by the following phase. At the same time, the following phase can also be assigned the following phase parameters of the main phase: green flash duration, yellow light duration, all red duration, pre-yellow duration. It should be noted that the main phase and the following phase can have different late start parameters and early termination parameters, so there is no need to assign the two parameters of the main phase to the following phase. Please continue to refer to Figure 2b , the following phase "South Right Turn" and the main phase "East Left Turn" are controlled synchronously. In addition, similar to the processing method of the main phase when there are multiple stage structures in the target data, if the same following phase is followed by multiple adjacent main phases in the stage structure, the start and end time of the main phase in the cycle length can be obtained by merging the phase duration of the following phase followed by the main phase in different stage structures. For example, if it is described in the signal control planning text that "South Right Turn follows East Left Turn and South Straight", the large language model will generate two key-value pairs (i.e. key-value pairs) of "South Right Turn" in the second structure, respectively recording that it follows "East Left Turn" and "South Straight", and if the two main phases are adjacent in stage (such as Figure 2bIf the "refPhase" of a traffic phase is not empty (e.g., "East Left Turn"), then the control parameters of the traffic phase in the cycle length can be obtained by merging the phase parameters of the following phases "East Left Turn" and "South Go Straight" respectively. In addition, if the "refPhase" of a traffic phase is empty (e.g., "default"), it indicates that the traffic phase needs to follow a main phase, but the specific main phase that the traffic phase needs to follow is not specified, so the default configuration needs to be used to determine the main phase that the traffic phase needs to follow. After determining the main phase that the traffic phase needs to follow, the control parameters of the traffic phase in the cycle length can refer to the specific process of obtaining the control parameters of the following phase with a non-empty "refPhase", which will not be repeated here. For example, in the example of the aforementioned "result2", the "refPhase" of the traffic phases "North Pedestrian", "South Pedestrian", "East Pedestrian", and "West Pedestrian" are all "default", so the default configuration can be used to determine the main phase that each of the pedestrian phases needs to follow. For example, it can be determined that the main phase that each of the pedestrian phases needs to follow is the motor vehicle straight phase adjacent to and parallel to the pedestrian phase, and then the control parameters of the pedestrian phase in the cycle length can be determined based on the phase parameters of the motor vehicle straight phase adjacent to and parallel to the pedestrian phase. For ease of understanding, please continue to refer to Figure 2b For example, the main phase that the pedestrian phase "North Pedestrian" needs to follow is "East Go Straight", the main phase that the pedestrian phase "South Pedestrian" needs to follow is "West Go Straight", the main phase that the pedestrian phase "East Pedestrian" needs to follow is "South Go Straight", and the main phase that the pedestrian phase "West Pedestrian" needs to follow is "North Go Straight". Of course, the above examples are only one possible example of the control parameters of the following phase and the pedestrian phase in actual application, and other possible cases will not be exemplified here.

[0118] In a specific implementation scenario, when processing the independent phase, the control parameters of the independent phase in the cycle length can be directly determined according to the phase parameters (e.g., the start time and the end time, or equivalently, the start time and the phase length) of the independent phase. In addition, similar to the processing method of the main phase when there are multiple stage structures, for the independent phase, if the same independent phase exists in multiple stages, the start and end time of the independent phase in the cycle length can be obtained by merging the phase lengths of the independent phase in different stages respectively.

[0119] In a specific implementation scenario, after determining the start and end time of each traffic phase in the cycle length, the specific value of the cycle length in the target data can be updated based on the start and end time of each traffic phase in the cycle length. Please continue to refer to Figure 2bIn this example, since the longest duration after the scheme assembly is determined to be 110 seconds after determining the start and end time of each traffic phase in the cycle duration, that is, the cycle duration specified in the third structure of the target data, the specific value of the cycle duration is still 110 seconds. Of course, Figure 2b The specific value of the cycle duration can be updated in this way under other possible circumstances, and examples are not repeated here.

[0120] In one implementation scenario, as a possible implementation example, after obtaining the signal control scheme of the target intersection through the foregoing flow steps, the signal control scheme can be directly issued, such as being issued to a signal machine, a signal control platform, etc., which is not limited here. For ease of understanding, please refer to Figure 2c , Figure 2c is a light color schematic diagram of an embodiment of the signal control scheme. As Figure 2c shown, in order to facilitate intuitive display of the signal control scheme, the signal control scheme is organized in the form of a light color diagram. Of course, other ways of organizing the signal control scheme in actual applications are not excluded, which are not limited here, and examples are not repeated here. In this example, the coordinates below the figure represent the corresponding time in the cycle duration, and each horizontal bar in the figure from 0 seconds to 110 seconds represents the control light color of the traffic phase shown on the left side of the horizontal bar at each time in the cycle duration, wherein green represents green light (dark green represents green light always on, light green represents green light flashing, yellow represents yellow light, and red represents red light). Of course, Figure 2c shown is only one possible example of the signal control scheme in the actual application process, and other possible circumstances are not limited here, and examples are not repeated here.

[0121] In another implementation scenario, as another possible implementation example, unlike the foregoing possible examples, after obtaining the signal control planning text, the target object can further input a signal control modification text to the signal control planning text at any time point, and in response to the signal control modification text of the target object to the signal control planning text, the context data of the interaction between the target object and the large language model can be obtained, and the context data can at least include the signal control planning text, and the signal control modification text modifies the jump change planned by the signal control planning text in natural language description. On this basis, a new large model instruction can be constructed based on the signal control modification text and the context data, and for the new large model instruction, the output data obtained by the large language model in response to the large model instruction can be returned as the target data. The steps of re-planning the signal control scheme, so as to support the target object to realize scheme modification.

[0122] In one specific implementation scenario, as mentioned above, the signal control modification text can also be described in natural language similar to the signal control planning text. For example, the signal control modification text can only modify the description of the part that needs to be modified in the signal control planning text, such as the signal control modification text can be partially modified in a description such as "the signal control planning about XXX in the above signal control planning text is modified to XXX"; or the signal control modification text can also be a comprehensive restatement of the signal control planning text, such as the signal control modification text can be a comprehensive restatement in a description such as "the above signal control planning text is adjusted to XXX". Of course, the above examples are only two possible examples of the signal control modification text in actual application, and the specific content of the signal control modification text is not limited here, and will not be exemplified one by one.

[0123] In one specific implementation scenario, the signal control modification text can specifically support but is not limited to the following modification types: modification of phase duration (including green light duration), adjustment of phase attribute value (such as adjustment of green flash time of straight phase), phase addition and deletion (such as addition of independent phase, deletion of following phase, etc.), change of phase or direction scheme structure (such as modification of phase scheme structure from stage type to ring type, or vice versa), change of phase following information (such as adjustment of the main phase followed by the following phase, etc.). Of course, the above examples are only a few possible examples of the modification types supported by the signal control modification text, and other modification types are not limited here, and will not be exemplified one by one.

[0124] In one specific implementation scenario, as one possible implementation example, the context data can include not only the signal control planning text, but also other content. For example, if the signal control modification text is obtained after the large language model parses the target data, the context data can also include the target data obtained by the large language model parsing the signal control planning text, so that the large language model can directly modify on the basis of the target data according to the signal control modification text, which helps to improve the parsing efficiency. In addition, the new large model instruction is reconstructed according to the signal control modification text and the context data, and its specific process can refer to the related description of constructing the large model instruction described above, which will not be repeated here. It needs to be emphasized that the main difference between the new large model instruction and the large model instruction constructed based on the signal control planning text is that the new large model instruction not only instructs the large language model to parse and output structured data about traffic phases, but also instructs the large language model to combine the reference signal control modification text and context data when parsing.

[0125] In yet another implementation scenario, as yet another possible implementation example, different from the foregoing possible examples, after the signal control scheme of the target intersection is generated, the scheme verification can also be performed based on the signal control scheme of the target intersection, and the scheme verification includes a plurality of verification items, and the plurality of verification items include at least one of the following: verifying whether the light group of the target intersection supports the signal control scheme, verifying whether the important phase of the target intersection has a phase service in the signal control scheme, and verifying whether there is a conflict in the signal control scheme. On this basis, in response to all the verification items being verified correctly, it can be determined that the signal control scheme is ready to be issued, and in response to at least one verification item being verified incorrectly, the target object can be prompted to modify the plan.

[0126] In a specific implementation scenario, in the process of verifying whether the light group of the target intersection supports the signal control scheme, for each traffic phase in the signal control scheme, the corresponding signal light of the target intersection can be checked respectively to determine whether the light group of the target intersection supports the signal control scheme. It should be noted that in signal control, the operation of some traffic phases requires a corresponding light group type. For example, the pedestrian phase must have a corresponding pedestrian signal light in the direction. For another example, in the foregoing example scheme, the independent operation of the south left-turn protection type phase requires that the target intersection must be installed with a left-turn direction indicating signal light (i.e., a left-turn arrow light) facing the south entrance; if the intersection only has a motor vehicle signal light (i.e., a disc light) facing the south entrance, the intersection cannot support the independent operation of the south left-turn phase.

[0127] In a specific implementation scenario, in the process of verifying whether the important phase of the target intersection has a phase service in the signal control scheme, the signal control scheme can be checked to determine whether there is a corresponding phase and light color setting for the important phase, and if not, it can be determined that there is a strong conflict phase release at the same time, that is, the verification fails. In addition, for non-important phases, such as left-turn, right-turn, and U-turn phases, no corresponding signal light or no corresponding phase (signal light on and off) can be allowed. It should be noted that, taking a cross intersection as an example, some traffic phases must have a corresponding signal light and phase service, such as the straight-through phase. Some traffic phases must have a corresponding phase if there is a corresponding signal light, such as the pedestrian phase. Traffic phases such as this can be referred to as important phases. Under certain circumstances, important phases can specifically refer to straight-through phases and pedestrian phases.

[0128] In a specific implementation scenario, in the process of checking whether there is a cross conflict in the signal control scheme, each pair of strong conflict phases in each second of the signal control scheme can be checked to see if at least one of them is red, and if not, it is determined that there is a strong conflict phase release at the same time, that is, the check fails. It should be noted that the strong conflict phase refers to a pair of flow directions that have cross conflicts at the target intersection and cannot be released at the same time. For example, north straight and east straight are a pair of strong conflict phases, and north straight and north pedestrian are a pair of strong conflict phases. In addition, for such as west straight and south right turn, it can be considered as a pair of weak conflict phases (not strong conflict phases), that is, it can be allowed to be released at the same time.

[0129] In a specific implementation scenario, as described above, if all the above checking items are checked without error, it can be determined that the signal control scheme is ready to be issued. For example, the scheme structure data can be generated according to the data protocol generation standard, and then issued to the signal machine, signal control platform, etc. For example, the signal control data protocol specified in “Road Traffic Signal Control Machine Information Release Interface Specification” GA / T1743-2020 can be referred to, which will not be repeated here. In addition, based on the phase light color result (such as the light color diagram shown in Figure 2c , the signal control scheme can be divided into several stages according to the start time and end time of each main phase, and then issued. For example, taking the signal control scheme shown in Figure 2c as an example, it can be divided into 5 stages (such as the white dotted line in Figure 2c ), which are: stage 1 (0-21 seconds) contains south left turn, north left turn and east right turn, stage 2 (21-39 seconds) contains north left turn, north straight and west pedestrian, stage 3 (39-65 seconds) contains north straight, west pedestrian, south straight and east pedestrian, stage 4 (65-85 seconds) contains east left turn, west left turn and south right turn, and stage 5 (85-110 seconds) contains east straight, west straight, north pedestrian, south pedestrian and east right turn, with a 3-second delay for east right turn. Of course, the above example is only one possible example of stage division in actual application, and other possible cases are not limited here, and will not be exemplified one by one.

[0130] The above scheme obtains the signal control planning text of the target object for the target intersection, and the signal control planning text is the jump change of the traffic signal planned by the target object in natural language. Then, based on the signal control planning text, a large model instruction is constructed, and the large model instruction is used to instruct the large language model to analyze the signal control planning text into structured data about the traffic phase. Thus, the output data of the large language model in response to the large model instruction is obtained as target data. Then, based on the target data, the signal control scheme of the target intersection is generated, and the signal control scheme includes the control parameters of each traffic phase in the cycle length. Therefore, on the one hand, the structured data is generated by analyzing the signal control planning text by the large language model, so as to generate the signal control scheme. Compared with manually inputting each parameter of each traffic phase, the generation efficiency of the signal control scheme can be improved. On the other hand, the target object can describe the signal control planning of the target intersection in natural language during the signal control planning process, so as to avoid performing a large number of manual operations, and to improve the operation convenience in the signal control scheme generation process as much as possible. Therefore, the generation efficiency of the signal control scheme can be improved, and the operation convenience in the signal control scheme generation process can be improved as much as possible.

[0131] Please refer to Figure 3 , Figure 3 is a schematic diagram of an embodiment of the signal control scheme generation device. The signal control scheme generation device 30 includes a planning acquisition module 31, an instruction construction module 32, a data acquisition module 33, and a scheme generation module 34. The planning acquisition module 31 is configured to obtain the signal control planning text of the target object for the target intersection. The signal control planning text is the jump change of the traffic signal planned by the target object in natural language. The instruction construction module 32 is configured to construct a large model instruction based on the signal control planning text. The large model instruction is used to instruct the large language model to analyze the signal control planning text into structured data about the traffic phase. The data acquisition module 33 is configured to obtain the output data of the large language model in response to the large model instruction as target data. The scheme generation module 34 is configured to generate the signal control scheme of the target intersection based on the target data. The signal control scheme includes the control parameters of each traffic phase in the cycle length.

[0132] The above scheme, the signal control scheme generation device 30 obtains the signal control planning text of the target object for the target intersection, and the signal control planning text is the jump change of the target object planning the traffic signal in natural language. Then, based on the signal control planning text, a large model instruction is constructed, and the large model instruction is used to instruct the large language model to analyze the signal control planning text into structured data about the traffic phase. Thus, the output data of the large language model responding to the large model instruction is obtained as the target data. Then, based on the target data, the signal control scheme of the target intersection is generated, and the signal control scheme includes the control parameters of each traffic phase in the cycle time. Therefore, on the one hand, the structured data is generated by analyzing the signal control planning text by the large language model to generate the signal control scheme, which can improve the generation efficiency of the signal control scheme compared with manually inputting each parameter of each traffic phase. On the other hand, the target object can describe the signal control planning of the target intersection in natural language during the signal control planning process, which can avoid performing a large number of manual operations and can improve the operation convenience as much as possible during the signal control scheme generation process. Therefore, the generation efficiency of the signal control scheme can be improved, and the operation convenience during the signal control scheme generation process can be improved as much as possible.

[0133] In some disclosed embodiments, the instruction construction module 32 includes a first construction submodule for constructing a prompt instruction template of the large language model based on a prompt word engineering. The prompt instruction template at least contains several reference signal control planning of different scheme structures, processing methods of different traffic phases, data output content and data output format of the large language model. The instruction construction module 32 includes a second construction submodule for constructing a large model instruction based on the prompt instruction template and the signal control planning text, so that the large language model analyzes the signal control planning text by referring to the prompt instruction template.

[0134] In some disclosed embodiments, the data acquisition module 33 includes a scheme acquisition submodule for acquiring, when the traffic phase in the signal control planning text involves a main phase, a scheme structure described by the large language model based on the main phase in the signal control planning text as a target structure. The scheme structure contains at least one of a ring type and a stage type. The data acquisition module 33 includes a structure creation submodule for acquiring a structure body created by the large language model for the target structure as a first structure body. The data acquisition module 33 includes a first update submodule for updating the first structure body based on the phase parameters at least containing the phase length described by the large language model in the signal control planning text to the main phase in the target structure. The target data contains the first structure body.

[0135] In some disclosed embodiments, the target data includes a second structure created by the large language model for recording phase attributes of the record, the data acquisition module 33 includes a first response submodule for updating the phase attributes of the main phase in the second structure based on the phase parameters of the main phase other than the phase duration parsed by the large language model in the signal control planning text in response to the large language model detecting that the traffic phase in the signal control planning text involves the main phase; the data acquisition module 33 includes a second response submodule for updating the phase attributes of the following phase in the second structure based on the main phase followed by the following phase parsed by the large language model in the signal control planning text in response to the large language model detecting that the traffic phase in the signal control planning text involves the following phase of the main phase; the data acquisition module 33 includes a third response submodule for updating the phase attributes of the independent phase in the second structure based on the start and end time of the independent phase parsed by the large language model in the signal control planning text in response to the large language model detecting that the traffic phase in the signal control planning text involves the independent phase independent of the main phase; the data acquisition module 33 includes a fourth response submodule for updating the phase attributes of the pedestrian phase in the second structure based on the phase parameters of the pedestrian phase parsed by the large language model in the signal control planning text in response to the large language model detecting that the traffic phase in the signal control planning text involves the pedestrian phase; wherein the phase parameters of the pedestrian phase include the main phase followed by the pedestrian phase specified in the signal control planning text, or the phase parameters of the pedestrian phase are a default placeholder when the main phase followed by the pedestrian phase is not specified in the signal control planning text.

[0136] In some disclosed embodiments, the target data includes a third structure created by the large language model for recording the cycle duration, the data acquisition module 33 includes a duration updating submodule for updating the third structure based on the cycle duration parsed by the large language model in response to the large language model detecting that the cycle duration is parsed in the signal control planning text; updating the third structure by the large language model with a target identifier in response to the large language model detecting that the cycle duration is not parsed in the signal control planning text; wherein the target identifier represents that the cycle duration is not specified in the signal control planning text.

[0137] In some disclosed embodiments, when the traffic phases in the signal control planning text involve main phases, follow-up phases of the main phases, and independent phases independent of the main phases, the scheme generation module 34 comprises a main phase processing submodule for determining the control parameters of each main phase in the cycle length based on the scheme structure of the main phase in the target data; wherein the scheme structure comprises at least one of a ring type and a stage type; the scheme generation module 34 comprises a target phase processing submodule for determining the control parameters of the target phase in the cycle length based on the control parameters of the main phase in the cycle length and the main phase followed by the target phase in the target data; and the scheme generation module 34 comprises an independent phase processing submodule for determining the control parameters of the independent phase in the cycle length based on the start and end time of the independent phase in the target data; wherein the target phase at least comprises a follow-up phase, and the control parameter at least comprises a start and end time.

[0138] In some disclosed embodiments, the main phase processing submodule is specifically configured to sequentially connect the phase lengths of the main phases in each scheme structure in the target data to obtain the start and end time of the main phase in the cycle length; and the target phase processing submodule is specifically configured to obtain the phase parameter of the main phase followed by the target phase in the target data, which is irrelevant to any of the late start and the early end, as a target parameter, and determine the control parameters of the target phase in the cycle length based on the target parameter and the phase parameter of the target phase in the target data, which is relevant to any of the late start and the early end.

[0139] In some disclosed embodiments, when there are multiple stage type structures in the target data, if the same main phase exists in multiple stage type structures, the start and end time of the main phase in the cycle length is determined by the phase lengths of the main phase defined in different stage type structures; and / or, when the target phase comprises a pedestrian phase and the phase parameter of the pedestrian phase in the target data is a default placeholder, the control parameters of the pedestrian phase in the cycle length are determined based on the phase parameters of the motor vehicle straight phase adjacent to and parallel with the pedestrian phase; and / or, when there are multiple stage type structures in the target data, if the same follow-up phase follows the main phase in multiple adjacent stage types, the start and end time of the main phase in the cycle length is obtained by merging the phase lengths of the main phase followed by the follow-up phase in different stage types.

[0140] In some disclosed embodiments, the signal control scheme generation device 30 comprises a cycle updating module for updating the specific value of the cycle length in the target data based on the start and end time of each traffic phase in the cycle length after determining the start and end time of each traffic phase in the cycle length.

[0141] In some disclosed embodiments, the signal control scheme generation apparatus 30 comprises a data acquisition module configured to acquire, in response to a signal control modification text of the target object to the signal control planning text, context data of an interaction between the target object and the large language model; wherein the context data at least comprises the signal control planning text, and the signal control modification text modifies the jump change planned by the signal control planning text in a natural language description; the signal control scheme generation apparatus 30 comprises a reconstruction module configured to reconstruct a new large model instruction based on the signal control modification text and the context data; the signal control scheme generation apparatus 30 comprises a loop iteration module configured to, for the new large model instruction, return to execute the step of acquiring output data of the large language model in response to the large model instruction as target data, so as to re-plan the signal control scheme.

[0142] In some disclosed embodiments, the signal control scheme generation apparatus 30 comprises a verification execution module configured to verify the signal control scheme based on the target intersection; wherein the scheme verification comprises a plurality of verification items, and the plurality of verification items comprise at least one of the following: verifying whether the light group of the target intersection supports the signal control scheme, verifying whether the important phase of the target intersection has a phase service in the signal control scheme, and verifying whether there is a conflict in the signal control scheme; the signal control scheme generation apparatus 30 comprises a determination and delivery module configured to determine to prepare to deliver the signal control scheme in response to all the verification items being verified correctly; the signal control scheme generation apparatus 30 comprises a prompt modification module configured to prompt the target object to modify the planning in response to at least one verification item being verified incorrectly.

[0143] In some disclosed embodiments, the interaction interface between the target object and the large language model displays a schematic diagram of the target intersection, and each traffic phase in the schematic diagram is marked with at least one of a phase name and a phase number, so as to be used by the target object to output the signal control planning text in a natural language by referring to the schematic diagram; and / or, the control parameters of the traffic phase in the cycle time length comprise a control light color of each time point of the traffic phase in the cycle time length.

[0144] Please refer to Figure 4 , Figure 4 is a frame schematic diagram of an embodiment of the electronic device of the present application. The electronic device 40 at least comprises a memory 41 and a processor 42 coupled with each other, the memory 41 at least stores program instructions, and the processor 42 is configured to execute the program instructions to realize the steps in any of the above signal control scheme generation method embodiments. For details, please refer to the foregoing disclosed embodiments, which will not be repeated here. In addition, as a possible example, the electronic device 40 can further comprise a display screen (not shown), which can be used to display content such as signal control planning text, signal control scheme, etc., to provide an interaction interface between the target object and the large language model. Of course, the electronic device 40 can also comprise other electronic elements (such as microphone, speaker, etc.), which are not limited here.

[0145] Specifically, the processor 42 is configured to control itself and the memory 41 to implement the steps in any of the above signal control scheme generation method embodiments. The processor 42 can also be referred to as a CPU (Central Processing Unit). The processor 42 can be an integrated circuit chip having a processing capability of signals. The processor 42 can also be a general processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component. The general processor can be a microprocessor or the processor can also be any conventional processor or the like. In addition, the processor 42 can be jointly implemented by integrated circuit chips.

[0146] In the above scheme, the electronic device 40 acquires a signal control planning text of a target object for a target intersection, and the signal control planning text plans a jump change of a traffic signal in a natural language for the target object, and then constructs a large model instruction based on the signal control planning text, and the large model instruction is used to instruct a large language model to analyze the signal control planning text into structured data about traffic phases, so as to acquire output data of the large language model responding to the large model instruction as target data, and then generate a signal control scheme of the target intersection based on the target data, and the signal control scheme includes control parameters of each traffic phase in a cycle time. Therefore, on the one hand, the structured data is generated by analyzing the signal control planning text by the large language model to generate the signal control scheme, which can improve the generation efficiency of the signal control scheme compared with manually inputting parameters of each traffic phase, and on the other hand, the target object can describe the signal control planning of the target intersection in a natural language during the signal control planning process, which can avoid performing a large amount of manual operations and can improve the operation convenience as much as possible during the signal control scheme generation process. Therefore, the generation efficiency of the signal control scheme can be improved, and the operation convenience during the signal control scheme generation process can be improved as much as possible.

[0147] Please refer to Figure 5 , Figure 5 is a frame diagram of an embodiment of the computer readable storage medium of the present application. The computer readable storage medium 50 stores program instructions 51 capable of being executed by a processor, and the program instructions 51 are used to implement the steps in any of the above signal control scheme generation method embodiments.

[0148] According to the scheme, the computer readable storage medium 50 acquires the signal control planning text of the target object for the target intersection, and the signal control planning text is a natural language planning of the jump change of the traffic signal by the target object. Then, based on the signal control planning text, a large model instruction is constructed, and the large model instruction is used to instruct the large language model to analyze the signal control planning text into structured data about the traffic phase, so as to acquire the output data of the large language model responding to the large model instruction as the target data, and then generate the signal control scheme of the target intersection based on the target data. The signal control scheme includes the control parameters of each traffic phase in the cycle time, so that on the one hand, the structured data is generated by analyzing the signal control planning text by the large language model to generate the signal control scheme, which can improve the generation efficiency of the signal control scheme compared with manually inputting each parameter of each traffic phase, and on the other hand, the target object can describe the signal control planning of the target intersection in natural language during the signal control planning process, which can avoid performing a large number of manual operations and can improve the operation convenience as much as possible during the signal control scheme generation process. Therefore, the generation efficiency of the signal control scheme can be improved, and the operation convenience during the signal control scheme generation process can be improved as much as possible.

[0149] In some embodiments, the device provided by the embodiments of the present disclosure has functions or includes modules that can be used to perform the methods described in the above method embodiments, and the specific implementation can refer to the description of the above method embodiments. For brevity, it will not be repeated here.

[0150] The above description of each embodiment tends to emphasize the differences between each embodiment, and the same or similar parts can be mutually referred to. For brevity, it will not be repeated here.

[0151] In several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the above-described device implementation is only schematic, for example, the division of the module or unit is only a logical function division, and in actual implementation, another division mode can be used, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed each other can be indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0152] The unit described as a separate component can or can not be physically separated, and the component shown as a unit can or can not be a physical unit, that is, it can be located in one place, or it can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the present embodiment scheme.

[0153] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0154] If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application, essentially or in the form of a contribution to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the methods in the various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program codes that can be stored in the medium.

[0155] If the technical solutions of the present application involve personal information, the product applying the technical solutions of the present application has been explicitly informed of the personal information processing rules before processing the personal information, and has obtained the personal independent consent. If the technical solutions of the present application involve sensitive personal information, the product applying the technical solutions of the present application has obtained the personal independent consent before processing the sensitive personal information, and at the same time meets the requirement of "explicit consent". For example, at the personal information collection device such as a camera, a clear and conspicuous sign is set to inform that the personal information collection range has been entered, and the personal information will be collected. If the individual voluntarily enters the collection range, it is considered to agree to collect the personal information. Or, on the device for processing personal information, the individual is informed of the personal information processing rules by using obvious signs / information, and the individual is authorized by means of pop-up information or by uploading the personal information by the individual himself. The personal information processing rules can include personal information processor, processing purpose, processing method, and processing personal information type, etc.

Claims

1. A method for generating a signal control scheme, characterized in that, The method comprises: obtaining a signal control planning text of a target object for a target intersection; wherein the signal control planning text is a natural language planning of the target object for the jump change of the traffic signal; based on the signal control planning text, constructing a large model instruction; wherein the large model instruction is used to instruct a large language model to analyze the signal control planning text into structured data about traffic phases; obtaining output data of the large language model in response to the large model instruction as target data; wherein the target data includes at least one of a first structure, a second structure, and a third structure, the first structure includes phase parameters of a main phase, a scheme structure composed of the main phase, and a phase order, the second structure includes phase parameters of each traffic phase, and the traffic phase includes at least one of a main phase, a following phase of the main phase, an independent phase independent of the main phase, and a pedestrian phase, and the third structure includes a cycle time length; based on the target data, generating a signal control scheme of the target intersection; wherein the signal control scheme includes control parameters of each traffic phase in a cycle time length.

2. The method of claim 1, wherein, The method comprises: based on prompt word engineering, constructing a prompt instruction template of the large language model; wherein the prompt instruction template includes at least several reference signal control plans of different scheme structures, processing methods of different traffic phases, data output content of the large language model, and data output format; based on the prompt instruction template and the signal control planning text, constructing the large model instruction for the large language model to analyze the signal control planning text based on the prompt instruction template.

3. The method of claim 1, wherein, When the traffic phase in the signal control planning text involves a main phase, the method of obtaining output data of the large language model in response to the large model instruction as target data comprises: obtaining a target structure created by the large language model based on a scheme structure of the main phase described in the signal control planning text; wherein the scheme structure includes at least one of a ring type and a stage type; obtaining a structure created by the large language model for the target structure as a first structure; updating the first structure based on phase parameters of the main phase described in the signal control planning text in the target structure analyzed by the large language model; wherein the target data includes the first structure.

4. The method of claim 1, wherein, The target data includes a second structure created by the large language model for recording phase attributes, and the method of obtaining output data of the large language model in response to the large model instruction as target data comprises at least one of: in response to the large language model detecting that the traffic phase in the signal control planning text involves a main phase, updating the phase attributes of the main phase in the second structure based on the phase parameters of the main phase other than the phase time length analyzed by the large language model in the signal control planning text. in response to the large language model detecting that the traffic phase in the signal control planning text involves a follow-up phase of a main phase, updating, based on the main phase that the follow-up phase follows that is parsed by the large language model in the signal control planning text, a phase attribute of the follow-up phase in the second structure; in response to the large language model detecting that the traffic phase in the signal control planning text involves an independent phase independent of a main phase, updating, based on the start and end time of the independent phase that is parsed by the large language model in the signal control planning text, a phase attribute of the independent phase in the second structure; in response to the large language model detecting that the traffic phase in the signal control planning text involves a pedestrian phase, updating, based on a phase parameter of the pedestrian phase that is parsed by the large language model in the signal control planning text, a phase attribute of the pedestrian phase in the second structure; wherein the phase parameter of the pedestrian phase includes a main phase that the pedestrian phase follows in the signal control planning text, or, when the main phase that the pedestrian phase follows is not specified in the signal control planning text, the phase parameter of the pedestrian phase is a default placeholder.

5. The method of claim 1, wherein, The target data includes a third structure created by the large language model for recording the cycle duration, and the output data of the large language model in response to the large model instruction is obtained as target data, including any one of the following: in response to the large language model detecting that the cycle duration is parsed in the signal control planning text, updating the third structure based on the cycle duration parsed by the large language model; in response to the large language model detecting that the cycle duration is not parsed in the signal control planning text, updating the third structure by the large language model with a target identifier; wherein the target identifier represents that the cycle duration is not specified in the signal control planning text.

6. The method of claim 1, wherein, When the traffic phase in the signal control planning text involves a main phase, a follow-up phase of the main phase, and an independent phase independent of the main phase, the generating the signal control scheme of the target intersection based on the target data includes: determining control parameters of each of the main phases in the cycle duration based on a scheme structure of the main phase in the target data; wherein the scheme structure includes at least one of a ring type and a stage type; determining control parameters of the target phase in the cycle duration based on the control parameters of the main phase in the cycle duration and the main phase that the target phase follows in the target data, and determining control parameters of the independent phase in the cycle duration based on the start and end time of the independent phase in the target data; wherein the target phase includes at least a follow-up phase, and the control parameters include at least start and end time.

7. The method of claim 6, wherein, The determining control parameters of each of the main phases in the cycle duration based on the scheme structure of the main phase in the target data includes: For each of the scheme structures in the target data, sequentially link based on the phase duration of the main phase in the scheme structure, to obtain the start and end time of the main phase in the cycle duration; And / or, the control parameters of the target phase in the cycle duration are determined based on the control parameters of the main phase in the cycle duration and the main phase followed by the target phase in the target data, including: Obtain the phase parameter of the main phase followed by the target phase in the target data, which is irrelevant to any of the late start and early end, as a target parameter, and determine the control parameters of the target phase in the cycle duration based on the target parameter and the phase parameter of the target phase in the target data, which is related to any of the late start and early end.

8. The method of claim 6, wherein, When there are multiple stage structures in the target data, if the same main phase exists in multiple stage structures, the start and end time of the main phase in the cycle duration is determined by the phase duration defined by the main phase in different stage structures respectively; And / or, when the target phase includes a pedestrian phase and the phase parameter of the pedestrian phase in the target data is a default placeholder, the control parameters of the pedestrian phase in the cycle duration are determined based on the phase parameters of the motor vehicle straight phase adjacent and parallel to the pedestrian phase; And / or, when there are multiple stage structures in the target data, if the same following phase follows the main phase in multiple adjacent stages, the start and end time of the main phase in the cycle duration is obtained by merging the phase duration of the following phase following the main phase in different stages.

9. The method of claim 6, wherein, After determining the start and end time of each traffic phase in the cycle duration, the method further comprises: Based on the start and end time of each traffic phase in the cycle duration, update the specific value of the cycle duration in the target data.

10. The method of claim 1, wherein, The method further comprises: In response to the signal control modification text of the target object to the signal control planning text, obtain the context data of the interaction between the target object and the large language model; wherein the context data at least includes the signal control planning text, and the signal control modification text modifies the jump change planned by the signal control planning text in natural language description; Based on the signal control modification text and the context data, construct a new large model instruction; For the new large model instruction, return to execute the step of obtaining the output data of the large language model in response to the large model instruction as target data, to re-plan the signal control scheme.

11. The method of claim 1, wherein, After generating the signal control scheme of the target intersection based on the target data, the method further comprises: Based on the signal control scheme of the target intersection, perform scheme verification; wherein the scheme verification includes a plurality of verification items, and the plurality of verification items include at least one of the following: verifying whether the light group of the target intersection supports the signal control scheme, verifying whether the important phase of the target intersection has a phase service in the signal control scheme, and verifying whether there is a conflict in the signal control scheme; In response to all of the check items being checked without error, it is determined that the signal control scheme is ready to be issued; In response to at least one of the check items being checked with error, the target object is prompted to modify the plan.

12. The method according to any one of claims 1 to 11, characterized in that, The interaction interface of the target object and the large language model displays a schematic diagram of the target intersection, and at least one of the phase name and the phase number is marked in each of the traffic phases in the schematic diagram, so that the target object can refer to the schematic diagram to output the signal control plan text in natural language; And / or, the control parameters of the traffic phases in the cycle length include the control light color of each of the traffic phases at each time in the cycle length.

13. A signal control scheme generation apparatus characterized by comprising: Comprise: The planning acquisition module is used for acquiring the signal control plan text of the target object for the target intersection; wherein, the signal control plan text is the jump change of the traffic signal planned by the target object in natural language; The instruction construction module is used for constructing a large model instruction based on the signal control plan text; wherein, the large model instruction is used for instructing the large language model to analyze the signal control plan text into structured data about traffic phases; The data acquisition module is used for acquiring the output data of the large language model in response to the large model instruction as target data; wherein, the target data includes at least one of a first structure, a second structure and a third structure, the first structure includes phase parameters of a main phase, a scheme structure composed of the main phase and a phase order, the second structure includes phase parameters of each of the traffic phases, the traffic phases include at least one of the main phase, a following phase of the main phase, an independent phase independent of the main phase and a pedestrian phase, and the third structure includes a cycle length; The scheme generation module is used for generating a signal control scheme of the target intersection based on the target data; wherein, the signal control scheme includes control parameters of each of the traffic phases in the cycle length.

14. An electronic device, comprising: At least comprising a memory and a processor, at least the memory stores program instructions, and the processor is used to execute the program instructions to realize the signal control scheme generation method of any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that, The program instructions capable of being run by the processor are stored, and the program instructions are used to realize the signal control scheme generation method of any one of claims 1 to 12.

Images