Traffic maintenance information processing method and device based on AI language large model, and storage medium

Abstract

The embodiment of the invention relates to the technical field of traffic, and discloses a traffic maintenance information processing method and device based on an AI language large model, and a storage medium, and the method comprises the steps: firstly obtaining traffic maintenance voice data uploaded by a user terminal, and converting the traffic maintenance voice data into text information through a voice recognition model; analyzing the text information to obtain the state characteristics of the target geographic position and the maintenance object, constructing a traffic congestion rate matrix according to the historical traffic state information of the target geographic position in different weather types and time periods, and extracting extreme value elements to obtain a first traffic congestion rate and a second traffic congestion rate; correcting the two according to the traffic congestion contribution degree corresponding to the state characteristics to obtain a first target traffic congestion rate and a second target traffic congestion rate, and triggering traffic maintenance work order distribution when the difference between the two is greater than or equal to a congestion rate threshold. According to the method, the influence condition of the state characteristics of the maintenance object on traffic congestion can be accurately evaluated, so that intelligent and timely distribution of the traffic maintenance work order is realized.

Description

Technical Field

[0001] The present invention relates to the field of traffic technology, and in particular to a traffic maintenance information processing method, device and storage medium based on an AI language large model. Background Art

[0003] Traditional traffic maintenance information processing relies primarily on manual collection and processing of various types of traffic maintenance information. For example, information on damage to traffic facilities and road construction progress is collected through reports from on-site inspectors and phone feedback from citizens. However, this approach has numerous drawbacks. Firstly, manual information collection is inefficient, making it difficult to comprehensively and comprehensively grasp the dynamic status of traffic maintenance in real time. This can lead to some traffic issues not being addressed promptly, causing congestion and even accidents. Secondly, manual information processing is prone to errors and omissions, and different personnel may interpret and record information differently, affecting the accuracy of traffic maintenance decisions. Summary of the Invention

[0004] The main purpose of the present invention is to provide a traffic maintenance information processing method, device and storage medium based on an AI language large model, aiming to solve the technical problems in the existing technology of low efficiency and easy omission and error in manual processing.

[0005] To achieve the above objectives, in a first aspect, an embodiment of the present application provides a method for processing traffic maintenance information based on an AI language large model, the method comprising:

[0006] Acquire traffic maintenance voice data uploaded by a user terminal, and convert the traffic maintenance voice data into text information through a pre-trained speech recognition model;

[0007] Parsing the geographic information and status information in the text information to obtain the target geographic location and status characteristics of the maintenance object, wherein the status characteristics include the damage type and its corresponding numerical parameters;

[0008] Constructing a traffic congestion rate matrix based on historical traffic status information of the target geographic location in different weather types and time periods;

[0009] Extracting extreme value elements from the traffic congestion rate matrix to obtain a first traffic congestion rate and a second traffic congestion rate;

[0010] Correcting the first traffic congestion rate and the second traffic congestion rate according to the traffic congestion contribution corresponding to the state feature to obtain a first target traffic congestion rate and a second target traffic congestion rate;

[0011] When the difference between the first target traffic congestion rate and the second target traffic congestion rate is greater than or equal to a congestion rate threshold, the dispatch of a traffic maintenance work order is triggered.

[0012] In a possible implementation, parsing the geographic information and status information in the text information to obtain the target geographic location and status characteristics of the maintenance object includes:

[0013] Identify geographic location description features through a pre-trained geographic name knowledge base and input them into a geocoding model to obtain the target geographic location;

[0014] The damage type is extracted through a pre-trained entity recognition model, and the descriptive size features are identified through a pre-trained size recognition model to obtain numerical parameters;

[0015] Inputting the damage type and corresponding numerical parameters into the traffic facility knowledge graph to obtain a verification matching degree;

[0016] The damage type and numerical parameter with a verification matching degree greater than or equal to the matching degree threshold are determined as the state characteristics of the maintenance object, and the damage type and numerical parameter with a verification matching degree less than the matching degree threshold are determined as invalid state characteristics.

[0017] In a possible implementation, constructing a traffic congestion rate matrix based on historical traffic status information of the target geographic location in different weather types and time periods includes:

[0018] Divide weather types into M levels and time periods into N segments;

[0019] Obtain traffic status information of the target geographic location within a preset historical time period, calculate the average congestion rate according to the weather-time combination, and construct an M*N traffic congestion rate matrix.

[0020] In a possible implementation, the correcting the first traffic congestion rate and the second traffic congestion rate according to the traffic congestion contribution corresponding to the state feature to obtain the first target traffic congestion rate and the second target traffic congestion rate includes:

[0021] According to the state characteristics of the maintenance object, the corresponding basic contribution coefficient is queried from the pre-built traffic congestion contribution knowledge base;

[0022] Calculate a traffic impact correction factor based on the road grade, number of lanes, and average daily traffic volume parameters of the target geographic location;

[0023] Multiply the basic contribution coefficient by the traffic impact correction coefficient to obtain the comprehensive traffic congestion contribution corresponding to the state characteristics;

[0024] Calculate the product of the first traffic congestion rate and the comprehensive traffic congestion contribution, and the product of the second traffic congestion rate and the comprehensive traffic congestion contribution respectively;

[0025] The first traffic congestion rate is added to the product of the first traffic congestion rate and the comprehensive traffic congestion contribution to obtain a first target traffic congestion rate;

[0026] The second traffic congestion rate is added to the product of the second traffic congestion rate and the comprehensive traffic congestion contribution to obtain the second target traffic congestion rate.

[0027] In one possible implementation, calculating the traffic impact correction coefficient based on the road grade, number of lanes, and average daily traffic volume parameters of the target geographic location includes:

[0028] The road grade, number of lanes, and average daily traffic volume parameters of the target geographic location are input into the traffic impact correction model to obtain a traffic impact correction coefficient. The traffic impact correction model satisfies the following expression:

[0029]

[0030] Among them, R is the road grade coefficient, L is the number of lanes, V is the average daily traffic volume, V0 is the standard traffic volume corresponding to the road grade, ω r is the road grade weight, ω l is the lane number weight, ω v is the daily average traffic flow weight, and α is the regional characteristic adjustment factor.

[0031] In a possible implementation, querying a corresponding basic contribution coefficient from a pre-built traffic congestion contribution knowledge base based on the state characteristics of the maintenance object includes:

[0032] The traffic congestion contribution knowledge base is hierarchically indexed, and the first contribution coefficient is obtained by matching the damage type of the maintenance object, and the second contribution coefficient is obtained by matching the severity of the state characteristics;

[0033] The first contribution coefficient and the second contribution coefficient are combined to obtain a basic contribution coefficient, wherein the combined contribution satisfies the following expression:

[0034] T=C t *C s *(1+θ), C t is the first contribution coefficient, C s is the second contribution coefficient, and θ is the expert evaluation adjustment factor.

[0035] In a possible implementation, extracting extreme elements from the traffic congestion rate matrix to obtain the first traffic congestion rate and the second traffic congestion rate includes:

[0036] Generate a congestion rate value ascending linked list from the elements in the traffic congestion rate matrix;

[0037] The first traffic congestion rate is obtained by extracting the tail node value in the ascending linked list of congestion rate values, and the second traffic congestion rate is obtained by extracting the head node value in the ascending linked list of congestion rate values.

[0038] In a possible implementation, triggering the dispatch of a traffic maintenance work order includes:

[0039] Obtaining current weather conditions, traffic congestion fluctuation rate, and maintenance urgency of the target geographic location;

[0040] Mapping the current weather conditions, traffic congestion volatility, and maintenance urgency into a three-dimensional decision space to obtain a comprehensive dispatch index;

[0041] The distribution mode is determined according to the comprehensive distribution index, wherein the distribution mode includes active distribution, autonomous order acceptance and mixed distribution.

[0042] In a second aspect, an embodiment of the present application further provides a traffic maintenance information processing device, comprising: a memory and a processor, wherein the memory is used to store program code; the processor is used to call the program code to execute the method described in the first aspect.

[0043] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method described in the first aspect when the computer program is executed by a processor.

[0044] Different from the existing technology, the traffic maintenance information processing method based on the AI ​​language large model provided in the embodiment of the present application first obtains the traffic maintenance voice data uploaded by the user terminal and converts it into text information using a voice recognition model. Then, the text is parsed to obtain the target geographic location and the status characteristics of the maintenance object (including damage type and numerical parameters). Then, a traffic congestion rate matrix is ​​constructed based on the historical traffic status information of the target geographic location in different weather types and time periods. Then, the extreme value elements are extracted to obtain the first traffic congestion rate and the second traffic congestion rate. Subsequently, the two are corrected according to the traffic congestion contribution corresponding to the status characteristics to obtain the first target traffic congestion rate and the second target traffic congestion rate. Finally, the traffic maintenance work order is triggered when the difference between the two is greater than or equal to the congestion rate threshold. The traffic maintenance information processing method of the present application effectively improves the efficiency and accuracy of traffic maintenance information processing, can accurately assess the impact of the status characteristics of the maintenance object on traffic congestion, thereby realizing the intelligent and timely dispatch of traffic maintenance work orders, helping to improve the intelligence level of urban traffic maintenance work and ensure the smooth operation of urban traffic. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0046] Figure 1 This is a flowchart of a traffic maintenance information processing method based on an AI language large model in some embodiments of the present application;

[0047] Figure 2 This is a flowchart of step S200 of the traffic maintenance information processing method based on the AI ​​language large model in some embodiments of the present application;

[0048] Figure 3 This is a flowchart of step S500 of the traffic maintenance information processing method based on the AI ​​language large model in some embodiments of the present application;

[0049] Figure 4 This is a schematic diagram of the hardware structure of the traffic maintenance information processing device in some embodiments of the present application.

[0050] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0052] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0053] In addition, the descriptions of "first", "second", etc. in the present invention are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0055] Traditional traffic maintenance information processing relies primarily on manual collection and processing of various types of traffic maintenance information. For example, information on damage to traffic facilities and road construction progress is collected through reports from on-site inspectors and phone feedback from citizens. However, this approach has numerous drawbacks. Firstly, manual information collection is inefficient, making it difficult to comprehensively and comprehensively grasp the dynamic status of traffic maintenance in real time. This can lead to some traffic issues not being addressed promptly, causing congestion and even accidents. Secondly, manual information processing is prone to errors and omissions, and different personnel may interpret and record information differently, affecting the accuracy of traffic maintenance decisions.

[0056] like Figure 1-Figure 3 As shown, the following takes the traffic maintenance information processing device executing the traffic maintenance information processing method based on the AI ​​language large model as an example for explanation. The traffic maintenance information processing device can be an edge server or a cloud server. It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a different order than here. Please refer to the attached Figure 1 The method includes the following steps S100 to S600:

[0057] Step S100: Acquire traffic maintenance voice data uploaded by a user terminal, and convert the traffic maintenance voice data into text information using a pre-trained speech recognition model;

[0058] Specifically, user terminals can be mobile phones, tablets, in-vehicle smart terminals, and other devices capable of uploading traffic maintenance voice data. This voice data can include descriptions of various traffic facility maintenance needs observed by users in real-world traffic scenarios, such as potholes, malfunctioning traffic lights, and damaged guardrails. After receiving the traffic maintenance voice data uploaded by the user terminal, a speech recognition model pre-trained with a large amount of traffic-related voice data is used to accurately recognize and convert the uploaded traffic maintenance voice data into text, converting it into structured text information for further analysis and processing.

[0059] Exemplarily, the cloud server obtains the traffic maintenance voice data uploaded by the user terminal, and converts the traffic maintenance voice data into text information through a pre-trained speech recognition model.

[0060] On the one hand, this application supports multiple user terminals uploading voice data at the same time, which greatly broadens the channels for information collection and enables traffic maintenance information in different scenarios to be timely aggregated into the system, improving the comprehensiveness and timeliness of information collection. On the other hand, with the help of a speech recognition model specially trained for the traffic field, it can accurately identify key information in the voice and efficiently convert voice data into text information, reducing the time and errors of manual transcription. This lays a solid foundation for the subsequent steps of quickly and accurately extracting key information, assessing traffic conditions, and triggering the dispatch of traffic maintenance work orders, thereby improving the efficiency and accuracy of the entire traffic maintenance information processing process, helping to promptly discover and resolve traffic maintenance problems and ensure the normal operation of the transportation system.

[0061] Step S200: parsing the geographic information and status information in the text information to obtain the target geographic location and status characteristics of the maintenance object, wherein the status characteristics include the damage type and its corresponding numerical parameters;

[0062] After obtaining text information through voice conversion, the system can use natural language processing technology to parse and identify key information in the text content.

[0063] For example, preset geographic information recognition rules and algorithms can be used to accurately locate and extract geographic information related to traffic maintenance from the text, such as specific road names, section numbers, intersection coordinates, etc., to determine the target geographic location. At the same time, semantic understanding and pattern matching technology are used to analyze the maintenance object status information described in the text to identify the type of damage to the maintenance object, such as cracks, potholes, and collapses in roads, rust and cracks in bridges, and damage and abnormal flashing of traffic lights. Numerical parameters corresponding to the damage type, such as the width and depth of cracks and the area and depth of potholes, are further extracted, thereby fully capturing the status characteristics of the maintenance object.

[0064] In one embodiment, the step S200 of parsing the geographic information and status information in the text message to obtain the target geographic location and status characteristics of the maintenance object includes:

[0065] S210, identifying geographic location description features through a pre-trained geographic name knowledge base and inputting them into a geocoding model to obtain a target geographic location;

[0066] S220, extracting the damage type using a pre-trained entity recognition model, and identifying descriptive size features using a pre-trained size recognition model to obtain numerical parameters;

[0067] S230, inputting the damage type and corresponding numerical parameters into the traffic facility knowledge graph to obtain a verification matching degree;

[0068] S240: Determine the damage type and numerical parameter whose verification matching degree is greater than or equal to the matching degree threshold as the state feature of the maintenance object, and determine the damage type and numerical parameter whose verification matching degree is less than the matching degree threshold as the invalid state feature.

[0069] Specifically, the embodiment of the present application first uses a pre-trained geographic name knowledge base (which includes various geographic names and their related feature information) to accurately identify the geographic location description features in the text information. After the geographic features are identified, they are input into the geocoding large model to quickly and accurately determine the target geographic location. Then, a pre-trained entity recognition model is used to accurately identify the types of damage described in the text, such as various types of damage to facilities such as roads, bridges, and traffic lights. At the same time, a pre-trained size recognition model is used, which focuses on identifying descriptive size features of damage in the text, such as length, width, depth, etc., and converts them into numerical parameters. The extracted damage types and corresponding numerical parameters are then input into the traffic facility knowledge graph. The traffic facility knowledge graph integrates rich traffic facility information, including common damage types and normal size ranges of various facilities. Through comparative analysis, the verification matching degree of these damage types and numerical parameters is obtained. Finally, the damage types and numerical parameters with a verification matching degree greater than or equal to the threshold are determined as valid status characteristics of the maintenance object. The valid status characteristics can accurately reflect the actual situation of the maintenance object; while the damage types and numerical parameters with a verification matching degree less than the threshold are determined as invalid status characteristics and are eliminated, and no subsequent work order dispatch analysis is performed.

[0070] This embodiment of the application utilizes pre-trained entity recognition and size recognition models to quickly and accurately extract damage types and corresponding numerical parameters from text, providing rich and detailed information for subsequent analysis. Furthermore, the extracted status features are verified using the traffic facility knowledge graph, effectively filtering out inaccurate or erroneous information, ensuring that the determined status features are authentic and reliable, and providing strong support for traffic maintenance decision-making.

[0071] Step S300: Constructing a traffic congestion rate matrix based on historical traffic status information of the target geographical location in different weather types and different time periods;

[0072] It is understood that in addition to the damage type of the maintenance object affecting traffic congestion rates, different weather types and different time periods also have a certain impact on traffic congestion. To comprehensively consider the factors affecting traffic congestion, this application first obtains historical traffic status information of the target location in different weather types and time periods after parsing the target location and status characteristics of the maintenance object. Then, a traffic congestion rate matrix is ​​constructed based on this historical traffic status information of the target location in different weather types and time periods.

[0073] In one embodiment, step S300 of constructing a traffic congestion rate matrix based on historical traffic status information of the target geographic location in different weather types and time periods includes:

[0074] Divide weather types into M levels and time periods into N segments;

[0075] Obtain traffic status information of the target geographic location within a preset historical time period, calculate the average congestion rate according to the weather-time combination, and construct an M*N traffic congestion rate matrix.

[0076] For example, weather types can be categorized as sunny / rainy / snowy / foggy, and time periods can be divided into morning peak / evening peak / off-peak / nighttime. Then, data from the target geographic location over the past three years is collected and the average congestion rate is calculated for each weather-time combination, forming a 4×4 matrix. For ease of presentation, the 4×4 matrix is ​​displayed in the following table.

[0077] Morning rush hour | Evening rush hour | Off-peak hours | Night time |

[0078] |----------|--------|--------|------|------|

[0079] | **Sunny** | 68% | 72% | 45% | 28% |

[0080] | **Rain** | 82% | 85% | 60% | 35% |

[0081] | **Snow** | 88% | 90% | 75% | 50% |

[0082] | **Fog** | 78% | 80% | 65% | 40% |

[0083] Step S400: extracting extreme elements from the traffic congestion rate matrix to obtain a first traffic congestion rate and a second traffic congestion rate;

[0084] Specifically, the elements in the traffic congestion rate matrix can be first generated into an ascending linked list of congestion rate values. The first traffic congestion rate is then obtained by extracting the tail node value from the ascending linked list, and the second traffic congestion rate is obtained by extracting the head node value from the ascending linked list. In other words, the first traffic congestion rate is the maximum traffic congestion rate for different weather types and time periods, and the second traffic congestion rate is the minimum traffic congestion rate for different weather types and time periods.

[0085] For example, the pre-built index L_asc is: 28% → 35% → 40% → 45% → 50% → 60% → 65% → 68% → 72% → 75% → 78% → 80% → 82% → 85% → 88% → 90%. Extraction is performed: First traffic congestion rate = L_desc header node value = 90% (snow / evening rush hour), Second traffic congestion rate = L_asc header node value = 28% (clear / nighttime).

[0086] Step S500: Correcting the first traffic congestion rate and the second traffic congestion rate according to the traffic congestion contribution corresponding to the state feature to obtain a first target traffic congestion rate and a second target traffic congestion rate;

[0087] It can be understood that the first traffic congestion rate and the second traffic congestion rate are the maximum and minimum values ​​of the average traffic congestion rates obtained based on historical traffic data statistics. If the status characteristics of the maintenance object have a greater impact on traffic, the first traffic congestion rate and the second traffic congestion rate need to be corrected according to the status characteristics of the maintenance object to obtain more accurate traffic congestion rates, thereby providing more accurate data support for whether emergency dispatch is needed in the future.

[0088] In one embodiment, step S500 of correcting the first traffic congestion rate and the second traffic congestion rate to obtain a first target traffic congestion rate and a second target traffic congestion rate according to the traffic congestion contribution corresponding to the state feature includes:

[0089] S510, querying a corresponding basic contribution coefficient from a pre-built traffic congestion contribution knowledge base according to the state characteristics of the maintenance object;

[0090] S520: Calculate a traffic impact correction coefficient based on the road grade, number of lanes, and average daily traffic volume parameters at the target geographic location;

[0091] S530: Multiply the basic contribution coefficient by the traffic impact correction coefficient to obtain a comprehensive traffic congestion contribution corresponding to the state characteristics;

[0092] S540, respectively calculating the product of the first traffic congestion rate and the comprehensive traffic congestion contribution, and the product of the second traffic congestion rate and the comprehensive traffic congestion contribution;

[0093] S550: Add the first traffic congestion rate to the product of the first traffic congestion rate and the comprehensive traffic congestion contribution to obtain a first target traffic congestion rate;

[0094] S560: Add the second traffic congestion rate to the product of the second traffic congestion rate and the comprehensive traffic congestion contribution to obtain a second target traffic congestion rate.

[0095] Different maintenance object status characteristics (such as cracked roads and faulty traffic lights) have varying degrees of impact on traffic congestion. By pre-building a knowledge base of traffic congestion contributions and assigning a corresponding basic contribution coefficient to each status characteristic, we can quantify the fundamental impact of these status characteristics on traffic congestion. For example, a large-scale road collapse has a much greater impact on traffic congestion than a slight flickering traffic light. The basic contribution coefficient can intuitively reflect this difference and provide a reasonable basis for subsequent corrections.

[0096] Parameters such as the road grade, number of lanes, and average daily traffic volume at the target location significantly influence the actual impact of state characteristics on traffic congestion. For example, on a high-volume highway with few lanes, a small pothole can cause severe traffic congestion; whereas, on a rural road with fewer lanes and less traffic, the same pothole may have less of an impact. By calculating a traffic impact correction coefficient that comprehensively considers these locational characteristics, we can more accurately reflect the actual impact of state characteristics on traffic congestion in a specific location, making the correction process more realistic.

[0097] This embodiment multiplies the basic contribution coefficient by the traffic impact correction coefficient to calculate the comprehensive traffic congestion contribution. This calculation method comprehensively considers the characteristics of the state characteristics themselves and the impact of geographic location characteristics on traffic congestion. By considering the common impact of different state characteristics in different scenarios and incorporating the individual factors of specific geographic locations, the calculated comprehensive traffic congestion contribution more comprehensively and accurately reflects the actual impact of state characteristics on traffic congestion, providing a reliable basis for subsequent correction of traffic congestion rates.

[0098] Finally, the embodiment of the present application calculates the product of the first traffic congestion rate and the comprehensive traffic congestion contribution, and the product of the second traffic congestion rate and the comprehensive traffic congestion contribution, and then adds these products to the original traffic congestion rate to obtain the first target traffic congestion rate and the second target traffic congestion rate. That is, the impact of the state characteristics on traffic congestion is superimposed on the original traffic congestion situation. Through this weighted correction method, the original traffic congestion rate can be reasonably adjusted to more accurately reflect the actual traffic congestion situation after considering the impact of the state characteristics. For example, if the state characteristics have a greater contribution to traffic congestion, then the corrected target traffic congestion rate will increase accordingly, which is more in line with the actual situation.

[0099] For example, the first target traffic congestion rate C target1 It is the product of the first traffic congestion rate C1 and the comprehensive traffic congestion contribution K (F), and the formula is: , ,in, is the traffic impact correction factor, is the basic contribution coefficient. The second target traffic congestion rate C target2 The calculation method of the first target traffic congestion rate C target1 The calculation method of is similar and will not be repeated here.

[0100] It can be seen that the traffic impact correction factor has a positive impact on the target traffic congestion rate. In other words, the larger the traffic impact correction factor, the higher the target traffic congestion rate. The traffic impact correction factor is related to parameters such as road grade, number of lanes, and average daily traffic volume. Therefore, the traffic impact correction factor can be quantified using road grade, number of lanes, and average daily traffic volume.

[0101] In one embodiment, the traffic impact correction coefficient is calculated based on the road grade, number of lanes, and average daily traffic volume parameters of the target geographic location, including:

[0102] The road grade, number of lanes, and average daily traffic volume parameters of the target geographic location are input into the traffic impact correction model to obtain the traffic impact correction coefficient β. The traffic impact correction model satisfies the following expression:

[0103]

[0104] Among them, R is the road grade coefficient, L is the number of lanes, V is the average daily traffic volume, V0 is the standard traffic volume corresponding to the road grade, ω r is the road grade weight, ω l is the lane number weight, ω v is the daily average traffic flow weight, and α is the regional characteristic adjustment factor.

[0105] Specifically, the road grade coefficient R is manually preset (e.g., expressway = 3, main road = 2, branch road = 1). is the weighted influence of road grade; the number of lanes L is the actual statistical value (such as 3 lanes, 5 lanes), is the weighted impact of the number of lanes (based on 3 lanes, only the extra lanes are included in the calculation to reflect the impact after exceeding the basic traffic capacity); the average daily traffic volume V is the actual statistical value (such as 10,000 vehicles / day), and the standard traffic volume V0 of the corresponding road grade is manually preset (such as the standard value of 50,000 vehicles / day for expressways). is the weighted impact of traffic flow. This application uses a regional characteristic adjustment factor α to further adjust the dynamic correction results to adapt to the traffic characteristics of different regions. The regional characteristic adjustment factor α is manually preset (e.g., 1.2 for the city center and 0.8 for the suburbs).

[0106] Step S600: When the difference between the first target traffic congestion rate and the second target traffic congestion rate is greater than or equal to a congestion rate threshold, triggering the dispatch of a traffic maintenance work order.

[0107] After correcting the first and second traffic congestion rates from historical data based on road grade, number of lanes, and average daily traffic volume to obtain first and second target traffic congestion rates, if the difference between the first and second target traffic congestion rates is greater than or equal to the congestion rate threshold, it indicates that the status characteristics of the maintenance object have a significant impact on traffic congestion at the target location, or that the status characteristics of the maintenance object have a significant impact on the traffic congestion rate in specific weather conditions or at specific time periods, deviating from the normal range. In this case, a traffic maintenance work order needs to be promptly triggered to perform maintenance on the maintenance object.

[0108] In one embodiment, the current weather conditions, traffic congestion volatility, and maintenance urgency of the target geographic location can be first obtained; then the current weather conditions, traffic congestion volatility, and maintenance urgency are mapped to a three-dimensional decision space to obtain a comprehensive distribution index; finally, a distribution mode is determined based on the comprehensive distribution index, wherein the distribution mode includes active distribution, autonomous order acceptance, and mixed distribution.

[0109] Current weather conditions significantly impact traffic conditions. For example, severe weather conditions such as heavy rain, snow, and fog can make roads slippery and reduce visibility, increasing the risk of accidents and worsening traffic congestion. Different weather conditions (such as sunny, cloudy, rainy, and snowy) can have varying impacts on the urgency and manner in which traffic maintenance work orders are dispatched. For example, during heavy rain, maintenance issues such as flooding and drainage failures may require more swift attention to ensure smooth and safe traffic flow.

[0110] The traffic congestion volatility reflects the changing levels of traffic congestion at a target location. By analyzing the traffic congestion volatility, we can determine whether traffic congestion is stable or fluctuating. A high traffic congestion volatility indicates unstable traffic conditions in the area, possibly due to unexpected factors that cause congestion to fluctuate. In such cases, timely addressing maintenance issues to stabilize traffic conditions is crucial. For example, if a road section frequently experiences congestion during peak hours, with varying degrees of congestion, analyzing the congestion volatility can determine whether prioritizing work orders for maintenance of the relevant facilities is necessary.

[0111] Maintenance urgency is determined based on the nature of the repair target, the extent of the damage, and the impact on traffic. For example, a serious problem like a large road collapse or a complete traffic light failure would have a high maintenance urgency and require immediate attention. Minor damage, such as flickering streetlights or blurred road markings, would have a lower maintenance urgency and can be addressed at an appropriate time.

[0112] This embodiment of the application constructs a three-dimensional decision space using the current weather conditions, traffic congestion volatility, and maintenance urgency as three dimensions. Each dimension has a different value range and quantitative standard. For example, weather conditions can be divided into different levels such as sunny, light rain, moderate rain, heavy rain, and rainstorm, and assigned corresponding numerical values; traffic congestion volatility can be calculated based on actual data to obtain a specific numerical range; and maintenance urgency can also be divided into three levels: low, medium, and high, each corresponding to a different numerical value.

[0113] The acquired values ​​for current weather conditions, traffic congestion fluctuations, and maintenance urgency are mapped into a three-dimensional decision space. A comprehensive dispatch index is then calculated using a specific algorithm or model. This index comprehensively considers the influencing factors of these three dimensions and can more comprehensively reflect the urgency and importance of dispatching maintenance work orders. For example, during heavy rain, when traffic congestion fluctuations are high and maintenance urgency is high, the calculated comprehensive dispatch index will be relatively high, indicating that the work order needs to be dispatched as soon as possible.

[0114] A high comprehensive dispatch index indicates a critical maintenance issue with a significant impact on traffic, necessitating proactive intervention by traffic management departments to promptly dispatch maintenance work orders to the appropriate personnel or teams. For example, if a rainstorm causes severe road flooding and traffic disruptions, traffic management departments should proactively contact the drainage maintenance team and quickly dispatch personnel to the site. This proactive dispatch model ensures that urgent maintenance issues are responded to and addressed as quickly as possible, minimizing traffic disruptions.

[0115] When the comprehensive dispatch index is low, it indicates that the maintenance issue is relatively low-urgency and has a minimal impact on traffic. In this case, a self-service order acceptance model can be adopted, where maintenance work orders are posted to a public platform or system, allowing maintenance personnel to independently choose whether to accept the order based on their own schedule and capabilities. For example, for minor facility damage, maintenance personnel can independently decide whether to address it based on their own availability and workload. This self-service order acceptance model can increase maintenance personnel's work enthusiasm and flexibility while also reducing management costs for traffic management departments.

[0116] When the comprehensive dispatch index is at a medium level, it indicates that the maintenance issue has a certain degree of urgency but does not require immediate proactive dispatch. In this case, a hybrid dispatch model can be adopted, whereby the traffic management department proactively dispatches some work orders to specific maintenance personnel or teams, while other work orders are posted to a public platform for maintenance personnel to accept on their own. For example, when traffic congestion fluctuates significantly but the weather and maintenance urgency are relatively normal, the traffic management department can proactively dispatch maintenance work orders for some key sections of road to professional maintenance teams, while posting work orders for some non-critical sections on the platform for other maintenance personnel to choose from. This hybrid dispatch model balances urgency and flexibility, improving the efficiency and quality of maintenance work.

[0117] Based on this, the traffic maintenance information processing method based on the AI ​​language large model provided in the embodiment of the present application first obtains the traffic maintenance voice data uploaded by the user terminal and converts it into text information using a voice recognition model. Then, the text is parsed to obtain the target geographic location and the status characteristics of the maintenance object (including damage type and numerical parameters). Then, a traffic congestion rate matrix is ​​constructed based on the historical traffic status information of the target geographic location in different weather types and time periods. Then, the extreme value elements are extracted to obtain the first traffic congestion rate and the second traffic congestion rate. Subsequently, the two are corrected according to the traffic congestion contribution corresponding to the status characteristics to obtain the first target traffic congestion rate and the second target traffic congestion rate. Finally, the traffic maintenance work order is triggered when the difference between the two is greater than or equal to the congestion rate threshold. The traffic maintenance information processing method of the present application effectively improves the efficiency and accuracy of traffic maintenance information processing, can accurately assess the impact of the status characteristics of the maintenance object on traffic congestion, thereby realizing the intelligent and timely dispatch of traffic maintenance work orders, helping to improve the intelligence level of urban traffic maintenance work and ensure the smooth operation of urban traffic.

[0118] like Figure 4 As shown, Figure 4 This is a hardware structure diagram of the traffic maintenance information processing device in some embodiments of the present application. The traffic maintenance information processing device provided in the embodiments of the present application includes a memory 1000 and a processor 2000, wherein the memory 1000 is used to store computer-readable instructions, and the processor 2000 is used to call the computer-readable instructions to execute the traffic maintenance information processing method based on the AI ​​language large model as described above.

[0119] Among them, the processor 2000 is used to provide computing and control capabilities to control the traffic maintenance information processing device to perform corresponding tasks, for example, controlling the traffic maintenance information processing device to perform the traffic maintenance information processing method based on the AI ​​language large model in any of the above-mentioned method embodiments, the method comprising: obtaining traffic maintenance voice data uploaded by a user terminal, and converting the traffic maintenance voice data into text information through a pre-trained voice recognition model; parsing geographic information and status information in the text information to obtain status characteristics of the target geographic location and the maintenance object, the status characteristics including the damage type and its corresponding numerical parameters; constructing a traffic congestion rate matrix based on historical traffic status information of the target geographic location in different weather types and different time periods; extracting extreme value elements from the traffic congestion rate matrix to obtain a first traffic congestion rate and a second traffic congestion rate; correcting the first traffic congestion rate and the second traffic congestion rate according to the traffic congestion contribution corresponding to the status characteristics to obtain a first target traffic congestion rate and a second target traffic congestion rate; and triggering the dispatch of a traffic maintenance work order when the difference between the first target traffic congestion rate and the second target traffic congestion rate is greater than or equal to a congestion rate threshold.

[0120] Processor 2000 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a hardware chip, or any combination thereof; it may also be a digital signal processing (DSP), an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or any combination thereof. The PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0121] Memory 1000, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the traffic maintenance information processing method based on the AI ​​language large model in the embodiments of this application. Processor 2000 can implement the traffic maintenance information processing method based on the AI ​​language large model in any of the aforementioned method embodiments by executing the non-transitory software programs, instructions, and modules stored in memory 1000.

[0122] Specifically, the memory 1000 may include a volatile memory (VM), such as a random access memory (RAM); the memory 1000 may also include a non-volatile memory (NVM), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD) or other non-transitory solid-state storage device; the memory 1000 may also include a combination of the above types of memory.

[0123] To sum up, the traffic maintenance information processing device of the present application adopts the technical solution of any one of the above-mentioned traffic maintenance information processing method embodiments based on the AI ​​language large model. Therefore, it has at least the beneficial effects brought by the technical solution of the above-mentioned embodiments, which will not be repeated here one by one.

[0124] The present application also provides a computer-readable storage medium, such as a memory device including program code. The program code can be executed by a processor to implement the traffic maintenance information processing method based on the AI ​​language large model described in the above embodiment. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CDROM), a magnetic tape, a floppy disk, or an optical data storage device.

[0125] The present application also provides a computer program product comprising one or more program codes stored in a computer-readable storage medium. A processor of the early warning system reads the program code from the computer-readable storage medium and executes the program code to complete the steps of the traffic maintenance information processing method based on the AI ​​language large model provided in the above-described embodiment.

[0126] Those skilled in the art will understand that all or part of the steps of implementing the above embodiments can be completed by hardware, or by hardware related to program code, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0127] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0128] Through the description of the above embodiments, it is clear to those skilled in the art that each embodiment can be implemented by means of software plus a general hardware platform, or of course by hardware. It is understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0129] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformation made by using the contents of the present description and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. A traffic maintenance information processing method based on AI language large model, characterized in that: The method comprises: Acquire traffic maintenance voice data uploaded by a user terminal, and convert the traffic maintenance voice data into text information through a pre-trained speech recognition model; Parsing the geographic information and status information in the text information to obtain the target geographic location and status characteristics of the maintenance object, wherein the status characteristics include the damage type and its corresponding numerical parameters; Constructing a traffic congestion rate matrix based on historical traffic status information of the target geographic location in different weather types and time periods; Extracting extreme value elements from the traffic congestion rate matrix to obtain a first traffic congestion rate and a second traffic congestion rate; Correcting the first traffic congestion rate and the second traffic congestion rate according to the traffic congestion contribution corresponding to the state feature to obtain a first target traffic congestion rate and a second target traffic congestion rate; When the difference between the first target traffic congestion rate and the second target traffic congestion rate is greater than or equal to a congestion rate threshold, the dispatch of a traffic maintenance work order is triggered.

2. The traffic maintenance information processing method based on the AI ​​language large model according to claim 1 is characterized in that: The step of parsing the geographic information and status information in the text information to obtain the target geographic location and status characteristics of the maintenance object includes: Identify geographic location description features through a pre-trained geographic name knowledge base and input them into a geocoding model to obtain the target geographic location; The damage type is extracted through a pre-trained entity recognition model, and the descriptive size features are identified through a pre-trained size recognition model to obtain numerical parameters; Inputting the damage type and corresponding numerical parameters into the traffic facility knowledge graph to obtain a verification matching degree; The damage type and numerical parameter with a verification matching degree greater than or equal to the matching degree threshold are determined as the state characteristics of the maintenance object, and the damage type and numerical parameter with a verification matching degree less than the matching degree threshold are determined as invalid state characteristics.

3. The traffic maintenance information processing method based on the AI ​​language large model according to claim 1 is characterized in that: The traffic congestion rate matrix is ​​constructed based on the historical traffic status information of the target geographical location in different weather types and different time periods, including: Divide weather types into M levels and time periods into N segments; Obtain traffic status information of the target geographic location within a preset historical time period, calculate the average congestion rate according to the weather-time combination, and construct an M*N traffic congestion rate matrix.

4. The traffic maintenance information processing method based on the AI ​​language large model according to claim 1 is characterized in that: The correcting the first traffic congestion rate and the second traffic congestion rate according to the traffic congestion contribution corresponding to the state feature to obtain a first target traffic congestion rate and a second target traffic congestion rate includes: According to the state characteristics of the maintenance object, the corresponding basic contribution coefficient is queried from the pre-built traffic congestion contribution knowledge base; Calculate a traffic impact correction factor based on the road grade, number of lanes, and average daily traffic volume parameters of the target geographic location; Multiply the basic contribution coefficient by the traffic impact correction coefficient to obtain the comprehensive traffic congestion contribution corresponding to the state characteristics; Calculate the product of the first traffic congestion rate and the comprehensive traffic congestion contribution, and the product of the second traffic congestion rate and the comprehensive traffic congestion contribution respectively; The first traffic congestion rate is added to the product of the first traffic congestion rate and the comprehensive traffic congestion contribution to obtain a first target traffic congestion rate; The second traffic congestion rate is added to the product of the second traffic congestion rate and the comprehensive traffic congestion contribution to obtain the second target traffic congestion rate.

5. The traffic maintenance information processing method based on the AI ​​language large model according to claim 4 is characterized in that: Calculating the traffic impact correction coefficient based on the road grade, number of lanes, and average daily traffic volume parameters of the target geographic location includes: The road grade, number of lanes, and average daily traffic volume parameters of the target geographic location are input into the traffic impact correction model to obtain a traffic impact correction coefficient. The traffic impact correction model satisfies the following expression: Among them, R is the road grade coefficient, L is the number of lanes, V is the average daily traffic volume, V0 is the standard traffic volume corresponding to the road grade, ω r is the road grade weight, ω l is the lane number weight, ω v is the daily average traffic flow weight, and α is the regional characteristic adjustment factor.

6. The traffic maintenance information processing method based on the AI ​​language large model according to claim 4 is characterized in that: The method of querying the corresponding basic contribution coefficient from a pre-built traffic congestion contribution knowledge base according to the state characteristics of the maintenance object includes: The traffic congestion contribution knowledge base is hierarchically indexed, and the first contribution coefficient is obtained by matching the damage type of the maintenance object, and the second contribution coefficient is obtained by matching the severity of the state characteristics; The first contribution coefficient and the second contribution coefficient are combined to obtain a basic contribution coefficient, wherein the combined contribution satisfies the following expression: T=C t *C s *(1+θ), C t is the first contribution coefficient, C s is the second contribution coefficient, and θ is the expert evaluation adjustment factor.

7. The traffic maintenance information processing method based on the AI ​​language large model according to claim 1 is characterized in that: The step of extracting extreme value elements from the traffic congestion rate matrix to obtain the first traffic congestion rate and the second traffic congestion rate includes: Generate a congestion rate value ascending linked list from the elements in the traffic congestion rate matrix; The first traffic congestion rate is obtained by extracting the tail node value in the ascending linked list of congestion rate values, and the second traffic congestion rate is obtained by extracting the head node value in the ascending linked list of congestion rate values.

8. The traffic maintenance information processing method based on the AI ​​language large model according to claim 1 is characterized in that: The triggering of traffic maintenance work order dispatch includes: Obtaining current weather conditions, traffic congestion fluctuation rate, and maintenance urgency of the target geographic location; Mapping the current weather conditions, traffic congestion volatility, and maintenance urgency into a three-dimensional decision space to obtain a comprehensive dispatch index; The distribution mode is determined according to the comprehensive distribution index, wherein the distribution mode includes active distribution, autonomous order acceptance and mixed distribution.

9. A traffic maintenance information processing device, characterized in that: include: A memory and a processor, wherein the memory is used to store program codes; The processor is configured to call the program code to execute the method according to any one of claims 1 to 8.

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

Images