Intelligent dynamic route optimization method and device based on AI and real-time traffic data

Abstract

The invention relates to the technical field of route dynamic optimization, in particular to an intelligent dynamic route optimization method, device and equipment based on AI and real-time traffic data and a computer storage medium. According to the method, the real-time traffic data, the power business requirements and the material characteristics are comprehensively utilized, the multi-target optimization model facing the power scene is constructed, and multiple targets such as transportation timeliness, comprehensive cost and safety are effectively balanced. Through a generalized adaptive A * algorithm, intelligent avoidance of special power constraints such as road load bearing, equipment quakeproof and over-limit limitation is realized, and path planning is dynamically adjusted based on real-time traffic trend prediction and time limit constraint. The method also has a self-adaptive optimization capability, can continuously adjust a weight strategy according to a historical transportation effect and business system feedback, and ensures that a planning scheme is always optimal. The electric power material transportation efficiency and safety are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of route dynamic optimization technology, and in particular to an intelligent dynamic route optimization method, apparatus, equipment, and computer storage medium based on AI and real-time traffic data. Background Technology

[0002] Against the backdrop of accelerated urbanization, traffic congestion is becoming increasingly severe, posing a significant challenge, especially to the transportation of power supplies. Traditional route planning methods, such as static navigation systems based on Dijkstra's or A* algorithms, rely primarily on fixed road network data and preset travel times, focusing solely on the shortest distance or time. They cannot adapt to real-time traffic conditions (such as traffic accidents, temporary traffic control, or severe weather) and fail to consider the specific needs of power transportation. Existing route planning technologies, including improved algorithms incorporating real-time congestion data in ordinary civilian scenarios, multi-node route planning to optimize delivery efficiency in logistics scenarios, and technologies emphasizing accident risk avoidance in hazardous materials transportation, all have significant limitations: they do not cover safety constraints such as bridge load-bearing verification, avoidance of oversized road sections, and vibration protection for precision equipment in the transportation of large power items; they do not incorporate cost factors specific to the power industry, such as oversized permit fees and equipment reinforcement costs; and they lack deep integration with power business systems (such as power plant inventory and emergency repair timelines). This results in the inability to achieve multi-objective optimization of timeliness, cost, and safety in scenarios such as emergency repairs and large-item transportation. This application aims to address the technical shortcomings of existing route planning technologies in the power material transportation scenario, such as lack of multi-objective dynamic optimization capabilities, failure to integrate real-time traffic data with power business needs, and inability to adaptively respond to emergencies and special transportation constraints. Summary of the Invention

[0003] Therefore, the technical problem to be solved by the present invention is to overcome the problems of existing path planning technology in the scenario of power material transportation, such as lack of multi-objective dynamic optimization capability, failure to integrate real-time traffic data and power business needs, and inability to adaptively respond to emergencies and special transportation constraints.

[0004] To address the aforementioned technical problems, this invention provides an intelligent dynamic route optimization method based on AI and real-time traffic data, comprising: Based on sensors, GPS devices, and mobile devices distributed throughout the city, raw real-time traffic datasets are generated by real-time monitoring of vehicle speed, traffic flow, driving direction, and user travel data. Standardized real-time traffic data is then generated through data cleaning, format conversion, standardization, and normalization methods. Based on standardized real-time traffic data, power business needs and material characteristics, a multi-objective path selection model for power scenarios is generated by constructing a comprehensive objective function and weight allocation method. Based on a multi-objective path selection model and standardized real-time traffic data, the optimal path node sequence is generated through the node comprehensive evaluation function, dynamic actual cost function and dynamic heuristic cost function of the generalized adaptive A* algorithm. Based on the optimal path node sequence, the final path recommendation is generated in the form of map, text or voice, and the traffic status is updated in real time. Based on historical transportation data, real-time traffic changes, and feedback from the power business system, an updated weight strategy is generated through a weight adaptive adjustment mechanism and fed back to the multi-objective path selection model.

[0005] Preferably, the method for generating the original real-time traffic dataset based on sensors, GPS devices, and mobile devices distributed throughout the city, by real-time monitoring of vehicle speed, traffic flow, driving direction, and user travel data, includes: Based on geomagnetic sensors, microwave sensors, and cameras, information on vehicle speed, traffic flow, and driving direction on the road is collected; Based on GPS devices, the system collects real-time location and driving trajectory data of vehicles via satellite positioning. Based on applications installed on mobile devices, collect users' travel behavior data; The aforementioned multi-source heterogeneous data are aggregated into the data acquisition layer to generate a raw real-time traffic dataset, which is then transmitted to the data processing layer.

[0006] Preferably, the process of generating standardized real-time traffic data through data cleaning, format conversion, standardization, and normalization methods includes: Based on the original real-time traffic dataset, data cleaning algorithms and rules are used to identify and remove noise, errors and duplicate information to generate accurate cleaned data. Based on the cleaned and accurate data, through format conversion processing, the speed data collected by different sensors are unified in units, and the GPS coordinates are converted into a format that the model can process, generating intermediate data with a unified format. Based on the uniformly formatted intermediate data, standardized real-time traffic data that can be used for model processing is generated through standardization and normalization operations. Based on standardized real-time traffic data, statistical analysis and data mining techniques are used to generate traffic flow trends and congested road segment distribution patterns.

[0007] Preferably, the step of generating a multi-objective path selection model for power scenarios based on standardized real-time traffic data, power business needs, and material characteristics, by constructing a comprehensive objective function and weight allocation method, includes: Based on the type of power transmission task, initial weights are set for timeliness, overall cost, and transportation safety; Based on the characteristics of power materials, the weight allocation is adjusted to generate a weight combination that is adapted to the characteristics of the materials. Based on the constraints of power business demand, adjust the weights to generate scenario-based weight strategies; By inputting a scenario-based weighting strategy, labeled traffic data, and special electricity costs into a comprehensive objective function, and balancing multiple objectives through a weighted summation method, a multi-objective path selection model is generated.

[0008] Preferably, the step of generating the optimal path node sequence based on a multi-objective path selection model and standardized real-time traffic data, using the node comprehensive evaluation function, dynamic actual cost function, and dynamic heuristic cost function of the generalized adaptive A* algorithm, includes: Based on the starting and ending points of power material transportation, the comprehensive evaluation value of each node is calculated through the node comprehensive evaluation function to generate a node priority sequence; Based on standardized real-time traffic data and the special constraints of electric power transportation, the dynamic actual cost from the starting point to each node is calculated through a dynamic actual cost function. Based on real-time traffic trend prediction and power transportation time constraints, the dynamic heuristic cost from each node to the destination is estimated through a dynamic heuristic cost function. By comprehensively evaluating all nodes, the optimal path node sequence is generated through a generalized adaptive A* algorithm search.

[0009] Preferably, the step of generating a final route recommendation based on the optimal path node sequence in the form of a map, text, or voice, and updating traffic status in real time includes: Based on the optimal path node sequence, a visual path is generated in a map navigation APP or in-vehicle navigation system through a map rendering engine. Based on a standardized real-time traffic data interface, the route refresh mechanism continuously updates information on changes in congested road sections and temporary road control measures. Based on user interaction feedback or path deviation warnings, the recommended path is dynamically adjusted through the path recalculation module. Generate the final route recommendation and continuously output it to the user in the form of map, text or voice.

[0010] Preferably, the step of generating an updated weight strategy and feeding it back to the multi-objective path selection model based on historical transportation data, real-time traffic changes, and feedback from the power business system, through a weight adaptive adjustment mechanism, includes: The weights are dynamically adjusted based on standardized real-time traffic data and emergency data. The weights are automatically adjusted based on historical transportation data feedback. Automatically switch weighting strategies based on linkage signals from the power business system; The updated weighting strategy is fed back into the multi-objective path selection model.

[0011] The present invention also provides an apparatus comprising: The data acquisition and preprocessing module is used to generate raw real-time traffic datasets based on sensors, GPS devices and mobile devices distributed throughout the city, by real-time monitoring of vehicle speed, traffic flow, driving direction and user travel data. It also generates standardized real-time traffic data through data cleaning, format conversion, standardization and normalization methods. The multi-objective path selection model construction module is used to generate a multi-objective path selection model for power scenarios based on standardized real-time traffic data, power business needs and material characteristics by constructing a comprehensive objective function and weight allocation method. The optimal path node sequence generation module is used to generate the optimal path node sequence based on a multi-objective path selection model and standardized real-time traffic data, through the node comprehensive evaluation function, dynamic actual cost function and dynamic heuristic cost function of the generalized adaptive A* algorithm. The final route recommendation generation module is used to generate final route recommendations based on the optimal route node sequence, in the form of maps, text, or voice, and update traffic status in real time. The multi-objective path selection model update module is used to generate an updated weight strategy based on historical transportation data, real-time traffic changes, and feedback from the power business system, and then feed it back to the multi-objective path selection model through a weight adaptive adjustment mechanism.

[0012] The present invention also provides an apparatus comprising: Memory, used to store computer programs; The processor is used to implement the steps of the above-described intelligent dynamic route optimization method based on AI and real-time traffic data when executing the computer program.

[0013] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described intelligent dynamic route optimization method based on AI and real-time traffic data.

[0014] The technical solution of the present invention has the following advantages compared with the prior art: The intelligent dynamic route optimization method based on AI and real-time traffic data described in this invention comprehensively utilizes real-time traffic data, power business needs, and material characteristics to construct a multi-objective optimization model for power scenarios, effectively balancing multiple objectives such as transportation timeliness, overall cost, and safety. Through the generalized adaptive A* algorithm, it achieves intelligent avoidance of special power constraints such as road load-bearing capacity, equipment vibration resistance, and overload restrictions, while dynamically adjusting route planning based on real-time traffic trend prediction and time constraints. The method also possesses adaptive optimization capabilities, continuously adjusting weight strategies based on historical transportation performance and business system feedback to ensure the optimal planning scheme is always achieved. This invention significantly improves the efficiency and safety of power material transportation, providing reliable guarantees for critical businesses such as emergency repairs and heavy-duty transport, while reducing overall operating costs through route optimization, supporting the intelligent transformation and upgrading of power transportation. Attached Figure Description

[0015] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the implementation of an intelligent dynamic route optimization method based on AI and real-time traffic data provided by this invention. Figure 2 This is a structural block diagram of an intelligent dynamic route optimization device based on AI and real-time traffic data provided in an embodiment of the present invention. Detailed Implementation

[0016] The core of this invention is to provide an intelligent dynamic route optimization method, device, equipment, and computer storage medium based on AI and real-time traffic data. This achieves synergistic optimization of the timeliness, overall cost, and transportation safety of power material transportation, effectively solving the technical defects of existing route planning technologies that are difficult to adapt to the special constraints and dynamic scenarios of power transportation.

[0017] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please refer to Figure 1. Figure 1 The flowchart illustrates the implementation of an intelligent dynamic route optimization method based on AI and real-time traffic data provided by this invention; the specific operation steps are as follows: S101: Based on sensors, GPS devices and mobile devices distributed throughout the city, the system generates raw real-time traffic datasets by monitoring vehicle speed, traffic flow, driving direction and user travel data in real time. Standardized real-time traffic data is then generated through data cleaning, format conversion, standardization and normalization methods. S102: Based on standardized real-time traffic data, power business needs and material characteristics, a multi-objective path selection model for power scenarios is generated by constructing a comprehensive objective function and weight allocation method. S103: Based on a multi-objective path selection model and standardized real-time traffic data, the optimal path node sequence is generated through the node comprehensive evaluation function, dynamic actual cost function and dynamic heuristic cost function of the generalized adaptive A* algorithm; S104: Based on the optimal path node sequence, generate the final path recommendation in the form of map, text or voice, and update the traffic status in real time; S105: Based on historical transportation data, real-time traffic changes, and feedback from the power business system, an updated weight strategy is generated through a weight adaptive adjustment mechanism and fed back to the multi-objective path selection model.

[0019] Based on the above embodiments, this embodiment will provide a detailed description of step S101: In some embodiments, a method for generating a raw real-time traffic dataset containing noise, errors, and duplicate information, based on geomagnetic sensors, microwave sensors, cameras, GPS devices, and mobile devices distributed on urban roads, by real-time monitoring of vehicle speed, traffic flow, driving direction, location trajectory, and user travel data, includes: Based on geomagnetic sensors, microwave sensors, and cameras, information on vehicle speed, traffic flow, and driving direction on the road is collected; Based on GPS devices, the system collects real-time location and driving trajectory data of vehicles via satellite positioning. Based on applications installed on mobile devices, collect users' travel behavior data; The above multi-source heterogeneous data are aggregated to generate the original real-time traffic dataset.

[0020] It's important to note that sensors, GPS devices, and mobile devices are all crucial sources of data collection. Geomagnetic sensors, microwave sensors, and cameras monitor vehicle speed, traffic flow, and direction of travel in real time on the road; GPS devices use satellite positioning to obtain the real-time location and trajectory of vehicles; and mobile devices collect user travel data through various applications. These data sources act like information collection stations distributed throughout the city, continuously gathering traffic data.

[0021] In some embodiments, based on the original real-time traffic dataset containing noise, errors, and duplicate information, data cleaning algorithms are used to identify and remove bad data, followed by preprocessing operations such as format conversion, standardization, and normalization to generate accurate and reliable standardized traffic data, including: Based on the original real-time traffic dataset, data cleaning algorithms and rules are used to identify and remove noise, errors and duplicate information to generate accurate cleaned data. Based on the cleaned and accurate data, through format conversion processing, the speed data collected by different sensors are unified in units, and the GPS coordinates are converted into a format that the model can process, generating intermediate data with a unified format. Based on the uniformly formatted intermediate data, standardized real-time traffic data that can be used for model processing is generated through standardization and normalization operations. Based on standardized real-time traffic data, statistical analysis and data mining techniques are used to generate traffic flow trends and congested road segment distribution patterns.

[0022] It's important to note that raw data often contains noise, errors, and duplicate information. The data cleaning process uses a series of algorithms and rules to identify and remove this problematic data, ensuring its accuracy and reliability. Preprocessing involves format conversion, standardization, and normalization to make the data suitable for subsequent model processing. For example, speed data collected from different sensors is standardized to the same unit, and GPS coordinate data is converted to a format suitable for model processing. In the data analysis phase, statistical analysis and data mining techniques are used to extract valuable information from massive amounts of data, such as traffic flow trends and the distribution patterns of congested road sections, providing strong data support for the model layer. Based on the above embodiments, this embodiment will provide a detailed description of step S102: In some embodiments, based on cleaned standardized traffic data, power business demands (such as substation commissioning deadlines, the "golden 4 hours" for fault repair, and power plant inventory data), and the characteristics of power materials (such as the vibration resistance requirements of precision equipment and the load-bearing limitations of large equipment), a multi-objective path selection model for power scenarios is generated by constructing a comprehensive objective function tailored to power demand and a weighted summation method. It should be noted that, in order to overcome the limitations of traditional single-objective power transmission route selection models, multi-objective route selection models for power scenarios have emerged. This model fully considers the core factors in the power transmission process, such as transportation timeliness (matching project nodes and emergency repair deadlines), comprehensive costs (including special transportation fees and equipment protection fees), and transportation safety (road compliance and equipment protection conditions). By constructing a comprehensive objective function that fits the power demand, it achieves a balance and optimization of multiple objectives. • Based on the type of power transportation task (such as emergency repair task, regular transportation task), timeliness (ω2) is set, and initial weights for timeliness, comprehensive cost and transportation safety are set; It should be noted that the weighting of "timeliness (ω2), overall cost (ω1), and transportation safety (ω3)" in the route selection model is based on the priority of the core needs in the power transportation scenario. Based on the urgency of the task and whether it relates to critical nodes in the power grid, priority is assigned: Emergency repair tasks (such as substation busbar faults and cable transportation after typhoons): Timeliness must be prioritized, as there is a "golden repair window" for power grid faults; delays can lead to widespread power outages and increased losses. Example: For "emergency cable transportation," timeliness is weighted at 0.6, and safety at 0.4, precisely because "repair timeliness directly affects power outage recovery time."

[0023] For routine transportation tasks (such as non-emergency delivery of wind turbine blades and transformers): there is no clear time limit for emergency repairs, and priority is given to ensuring "safety" (avoiding equipment damage) and "overall cost" (controlling special transportation expenses). For example, in the "routine transportation of wind turbine blades", safety is weighted at 0.5, cost at 0.3, and timeliness at only 0.2, based on the fact that "the repair cost of blade bending damage is far higher than the cost of extended transportation time".

[0024] Specifically, when considering transportation timeliness, the model combines real-time traffic data (such as highway congestion and temporary traffic control information on national highways) with power service demands (such as substation commissioning deadlines and the "golden 4 hours" for fault repair) to accurately calculate transportation time for different road sections. This includes information such as delays caused by current congestion and the permitted travel periods for heavy equipment transportation permits, thereby predicting the entire material transportation cycle. For example, when repairing cables after a typhoon, the model can avoid flooded and congested road sections in real time, ensuring that materials are delivered within the repair window. Comprehensive cost factors cover expenditures unique to power transportation: fuel costs (for heavy transport vehicles), oversized vehicle transportation permit fees, road reinforcement fees (some road sections require temporary reinforcement to support overweight equipment), equipment transportation insurance fees, and parking monitoring fees (precision equipment requires dedicated personnel to be on duty at night). Transportation safety and equipment protection are relatively complex and require evaluation based on multiple indicators: road smoothness (affecting the shock resistance of precision equipment), bridge / tunnel load-bearing capacity (matching the weight of large equipment), traffic volume (avoiding frequent sudden braking that could scratch oversized equipment), and the presence of road sections prone to severe weather (such as mountainous areas prone to blizzards, which could lead to transportation delays).

[0025] For example, the route for transporting photovoltaic inverters, although shorter, passes through the city center (with heavy traffic and frequent starts and stops), which may cause internal components of the inverters to loosen due to bumps; while another slightly longer ring road route has less traffic and a smoother road surface, although it takes 30 minutes longer, it can significantly reduce the risk of equipment damage. It should be noted that the quantification of safety costs in the example is based on "meeting the safety requirements for the transportation of electrical materials (such as emergency cables)," and adopts the logic of "multi-indicator 10-point scoring + normalized conversion," specifically defined as follows: (1) Determine the core indicators for safety assessment: Focus on the key dimensions that affect the safety of cable transportation, including road surface smoothness (avoiding insulation wear), road section load-bearing / height limit compliance (eliminating the risk of exceeding limits), real-time traffic flow (reducing the impact of scrapes / sudden braking), and the impact of severe weather (avoiding hidden dangers such as slippage).

[0026] (2). 10-point scoring system: The score is based on the rule of "the lower the safety risk, the higher the score". For example, newly built high-grade highways (smoothness 10 points), bumpy road sections (low score), fully compliant road sections (10 points), low traffic volume road sections (high score), no severe weather (10 points).

[0027] (3) Normalization Conversion: In order to unify the calculation dimension with "timeliness (hours)", the 10-point score is converted into a coefficient of 0 to 1 (score ÷ 10). The higher the coefficient, the lower the safety risk and the lower the safety cost. For example, route B (newly built highway) scores 8 points → converted to 0.8 ("0.8 is a high score"), route A (bumpy road section) scores 3.5 points → converted to 0.35. The final coefficient is substituted into the comprehensive cost weighting calculation.

[0028] • Based on the characteristics of power materials (such as precision equipment relay protection devices, large heavy objects such as 300-ton transformers), adjust the weight allocation to generate a weight combination that matches the characteristics of the materials. It should be noted that: The "damage risk" and "transportation sensitivity" of different materials determine the weighting: Precision equipment (such as relay protection devices and photovoltaic inverters) is sensitive to "bumps and vibrations". Even minor damage may cause equipment failure. Therefore, the weight of "transportation safety" needs to be increased (e.g., ω3≥0.4 during regular transportation). Large and heavy objects (such as 300-ton transformers and 40-meter wind turbine blades): It is necessary to prioritize avoiding the risks of "over-limit and insufficient load-bearing capacity", and the special transportation cost accounts for a high proportion. Therefore, the weights of "safety" (ω3) and "comprehensive cost" (ω1) should be higher than those of ordinary materials. For conventional materials (such as cables and coal): the risk of damage is low and the transportation cost is controllable, so the weight of "safety" can be appropriately reduced, and the focus can be on "timeliness" or "cost" (for example, the transportation of coal needs to balance "power plant inventory time limit" and "transportation cost").

[0029] • Based on the constraints of power business demand (such as project node constraints, cost control constraints, and safety compliance constraints), adjust the weights to generate scenario-based weight strategies; It should be noted that: The weights must be aligned with the hard constraints of power grid engineering and power plant operation: Project milestone constraints (such as substation expansion deadlines): If materials need to be delivered before a fixed date to ensure grid connection, the weight of "timeliness" needs to be increased (e.g., ω2=0.5~0.7). Cost control constraints (such as the annual transportation budget of power companies): When transporting routine materials, if the budget is limited, the weight of "comprehensive cost" needs to be increased (e.g., ω1=0.4~0.5) to avoid over-limit fines and overspending on equipment reinforcement costs; Safety and compliance constraints (such as power grid safety regulations): The weight of "safety" in all transportation tasks shall not be lower than 0.2 (minimum 0.2) to ensure that no illegal transportation (such as passing through sections of road with insufficient load-bearing capacity) is caused by too low a weight.

[0030] • Input the scenario-based weighting strategy, along with labeled traffic data and special electricity costs (such as special transportation fees and equipment protection fees), into the comprehensive objective function, and balance multiple objectives through a weighted summation method to generate a multi-objective path selection model.

[0031] It should be noted that, in order to find a balance among multiple objectives, multi-objective path selection models for power scenarios typically employ a weighted summation method. This method assigns a weight to each objective, the magnitude of which reflects the importance of that objective to the power transmission task.

[0032] For example, for emergency substation repairs (such as transporting a backup busbar due to a busbar failure), the timeliness weight might be set to 0.7 (delivery must be completed within 4 hours), and the transportation safety weight to 0.3 (avoiding busbar collisions). For routine wind turbine blade transportation (without urgent project milestones), the transportation safety weight might be set to 0.5 (avoiding blade bending), the overall cost weight to 0.3 (controlling special transportation fees), and the timeliness weight to 0.2 (delivery within 3 days is sufficient). By adjusting these weights, the model can output the optimal route for different scenarios, meeting the personalized transportation needs of power companies.

[0033] In one specific embodiment, suppose a power transportation company needs to transport a batch of emergency cables to a faulty substation. Both timeliness and transportation safety are critical considerations in this task. The model sets the weight for timeliness to 0.6 and the weight for transportation safety to 0.4. When calculating the route, the model combines real-time traffic data (such as congested highway sections and national highway traffic conditions) with road safety parameters (such as road surface smoothness and the presence of height restrictions) to calculate the "time cost" (e.g., route A takes 2.5 hours but passes through one congested point, while route B takes 3 hours but is entirely unobstructed) and "safety cost" (e.g., route A has two bumpy sections, while route B is a newly built high-grade highway). The weighted sums are then used to obtain the comprehensive cost. Ultimately, if the overall cost of route B (0.6×3 + 0.4×0.8 = 2.12, with safety costs converted to a 10-point scale, where 0.8 is the highest score) is lower than that of route A (0.6×2.5 + 0.4×3.5 = 2.3), then route B will be selected as the optimal route and recommended to the transportation dispatchers, ensuring timely cable delivery while avoiding equipment damage. Based on the above embodiments, this embodiment will provide a detailed description of step S103: It should be noted that traffic conditions change rapidly in real-time environments, making it difficult for traditional path optimization algorithms to meet the demands for efficient and accurate path planning. To address this challenge, hybrid dynamic path optimization algorithms have emerged, integrating various advanced algorithmic ideas and optimization strategies to achieve more intelligent and flexible path planning. The generalized adaptive A* algorithm, based on the classic A* algorithm, adds a dynamic adjustment factor (addressing real-time traffic and scenario constraints). Its core formula consists of three parts: a "node comprehensive evaluation function," an "actual cost function," and a "dynamic heuristic function."

[0034] In some embodiments, based on a multi-objective path selection model and standardized real-time traffic data, the optimal path node sequence is generated using the node comprehensive evaluation function, dynamic actual cost function, and dynamic heuristic cost function of the generalized adaptive A* algorithm, including: Based on the starting and ending points of power material transportation, the comprehensive evaluation value of each node is calculated through the node comprehensive evaluation function, and a node priority sequence is generated (the smaller the value, the higher the priority). It should be noted that the node comprehensive evaluation function (core decision formula) is used to determine the priority of the current node (a transit point in the transportation path, such as a highway entrance / exit or a goods transfer station). The formula is as follows: f adapt (n) = g adapt (n) + h adapt (n) f adapt(n) : The comprehensive evaluation value of node n in the generalized adaptive A* algorithm (the smaller the value, the higher the priority of the node, and the more likely it is to be included in the path search). g adapt (n) The dynamic actual cost from the transportation origin S (such as an equipment factory or coal mine) to node n (including the comprehensive cost of real-time traffic and power transportation constraints). h adapt (n) The dynamic heuristic cost from node n to the transportation destination T (such as a substation or wind farm) (an estimated cost corrected based on real-time data, rather than a fixed distance). • Based on standardized real-time traffic data and special constraints of electric power transportation, the dynamic actual cost from the starting point to each node is calculated through a dynamic actual cost function; It should be noted that the dynamic actual cost function g adapt (n) (Core Improvement 1: Real-time Adjustment of Actual Costs) The classical A's g(n) only calculates the "physical distance from the starting point to node n", while the generalized adaptive A's g adapt (n) Real-time traffic costs and special costs of electricity transportation need to be combined, and the formula is:

[0035] Symbol decomposition and adaptation to power scenarios: d(k,k+1): The physical distance of road segment (k,k+1) (unit: km, e.g., the distance from "equipment factory to A highway entrance"); C cong (k,k+1): Real-time congestion coefficient of road segment (k,k+1) (dynamic value: C when traffic is smooth) cong =1, light congestion =1.2~1.5, heavy congestion =2~3, data comes from real-time interface of the transportation department or feedback from vehicle GPS; Example of a power scenario: When transporting wind turbine blades, if the route passes through suburban sections with traffic congestion due to construction (C cong If the cost of the road segment is 2.5, the actual cost of that road segment will double, and the algorithm will tend to avoid it. C spec (k,k+1): Special cost coefficient for power transportation in segment (k,k+1) (static + dynamic coefficient adapted to the characteristics of power materials, with a value ≥1). Core sub-items (specifically for power scenarios): C load (k,k+1): Road load-bearing capacity adaptability coefficient (e.g., when transporting a 300-ton transformer, if the bridge's load-bearing capacity is only 200 tons, C). load=+∞, directly exclude this road section; C when the load-bearing capacity meets the standard. load =1); C shock (k,k+1): Vibration resistance coefficient of precision equipment (when transporting relay protection devices, C on bumpy road sections) shock =1.8, smooth highway =1); C limit (k,k+1): Oversized component adaptation coefficient (when transporting a 40-meter wind turbine blade through a tunnel with a length limit of 30 meters, C) limit =+∞, exclude this road segment); Computational logic: C spec (k,k+1) = max{C load C shock C limit (Use the most stringent constraint coefficient to ensure transportation compliance and safety); α, β: Weighting coefficients (α+β=1, α=0.6 for emergency power transportation (prioritize avoiding congestion and ensuring timeliness), β=0.7 for regular large-item transportation (prioritize ensuring safety and compliance)).

[0036] • Based on real-time traffic trend prediction and power transportation time constraints, the dynamic heuristic cost from each node to the destination is estimated through a dynamic heuristic cost function; It should be noted that h(n) in classical A is often the "straight-line distance from node n to the endpoint T" (a fixed value), while h in generalized adaptive A is... adapt (n) needs to be dynamically adjusted based on real-time traffic trends and power transmission endpoint constraints (such as project deadlines), as shown in the formula: H adapt (n) = h base (n,T)*γC trend (n,T)*C deadline (T) Symbol decomposition and adaptation to power scenarios: h base (n,T): The basic estimated cost from node n to destination T (using the "shortest physical distance" or "historical best transportation cost" as the benchmark value); C trend (n,T): Real-time traffic trend coefficient for the path from n to T (based on traffic forecasts for the next 1-2 hours; if congestion is predicted to worsen, C...). trend =1.3~1.5; if the forecast is smooth, C trend =0.9~1.0 (data from AI traffic prediction model). Example of a power scenario: Transporting emergency cables to a faulty substation. If it is predicted that the roads around the destination will be congested during the evening rush hour one hour later (C... trendIf the value is 1.4, the heuristic cost will increase, and the algorithm will tend to choose the path that is "a little further away now but not congested in the future"; C deadline (T): Time constraint coefficient for the power transmission endpoint T (adaptation project / emergency repair time: if T is an emergency repair point and the remaining time is <2 hours, C) deadline =1.6 (priority time limit); if T is a conventional wind farm with no emergency time limit, C deadline =1.0); γ: Heuristic weighting coefficient (usually taken as 0.8~1.2, balancing "estimation accuracy" and "search efficiency". For the transportation of large power items, γ=1.1 to avoid missing compliant paths due to estimation bias).

[0037] Note: The symbol definitions of the three core formulas involved in the generalized adaptive A* algorithm—the node comprehensive evaluation function, the dynamic actual cost function, and the dynamic heuristic cost function—are clarified in conjunction with the power transportation scenario, specifying their value ranges and units. They are categorized by function module as shown in Table 1. Table 1 Symbol Definitions

[0038] • Evaluate all nodes comprehensively and generate the optimal path node sequence through a generalized adaptive A* algorithm search.

[0039] It should be noted that, Machine learning algorithms form the foundation of AI applications in the transportation sector. By learning from and analyzing vast amounts of historical traffic data, they uncover hidden patterns and rules, thereby establishing traffic prediction and route optimization models. Taking traffic flow prediction as an example, commonly used machine learning algorithms such as linear regression, decision trees, and support vector machines can build predictive models based on historical traffic flow data, time, date, weather, and other factors. These models can analyze the degree of influence of different factors on traffic flow and predict future traffic flow trends based on current conditions. For instance, by analyzing daily morning rush hour traffic flow data for a specific road segment over the past year, along with corresponding weather and date information, and using a linear regression algorithm to build a model, when the current date and weather information are input, the model can predict the approximate traffic flow for that road segment during the morning rush hour that day. As a branch of machine learning, deep learning algorithms have demonstrated strong application potential in the transportation field in recent years. Deep learning algorithms possess powerful automatic feature extraction capabilities and the ability to model complex nonlinear relationships, enabling them to handle more complex traffic data and scenarios. In the transportation field, recurrent neural networks (RNNs) and their variants, such as long short-term memory networks (LSTMs) and gated recurrent units (GRUs), are commonly used to process traffic data with time-series characteristics, such as traffic flow changes over time. These algorithms can effectively capture long-term dependencies in traffic data and accurately predict short-term and long-term trends in traffic flow. For example, by using an LSTM network to model traffic flow at multiple intersections in a city, and by learning the time-series features from historical data, the model can predict peak and off-peak traffic periods at different intersections hours or even days in advance, providing a basis for traffic management departments to formulate traffic control plans in advance. In terms of route selection, AI utilizes classic path search algorithms such as Dijkstra's algorithm and A* algorithm, combined with real-time traffic data and prediction results, to plan the optimal route for users. These algorithms are based on graph theory, abstracting the road network as a graph composed of nodes (intersections) and edges (road segments). By calculating the shortest or optimal path from the starting point to the destination, they determine the best travel route. In practical applications, AI acquires real-time information such as road congestion and traffic speed, incorporating these factors into the cost function of path calculation. For example, when a road segment is congested, the algorithm automatically increases the toll cost for that segment, guiding the path search to avoid congested sections and choose smoother routes. Simultaneously, AI also considers users' personalized needs, such as travel preferences (whether they prefer highways, avoid toll roads, etc.), and travel time constraints, providing users with personalized optimal route planning.

[0040] Beyond traffic flow prediction and route selection, AI plays a crucial role in traffic signal control and traffic accident prediction. In traffic signal control, AI can dynamically adjust traffic light timing schemes based on real-time traffic flow, improving intersection efficiency and reducing vehicle waiting time. In traffic accident prediction, AI analyzes historical accident data, traffic flow, road conditions, weather, and other multi-source information to build accident prediction models, predicting potential accident locations and times in advance, providing decision support for traffic management departments to take preventative measures.

[0041] It's important to note that multi-objective route selection models calculate the optimal route based on the traveler's origin, destination, and real-time traffic data. For example, these models comprehensively consider factors such as time, cost, and comfort to provide users with more personalized and comprehensive route planning. They also utilize traffic prediction models, such as machine learning and deep learning models, to predict future traffic conditions, allowing for better route planning in advance. These models work together, continuously adjusting and optimizing route planning based on real-time traffic data and prediction results to ensure that the routes provided to users are always optimal. Based on the above embodiments, this embodiment will provide a detailed description of step S104: In some embodiments, based on the optimal path node sequence, the final route recommendation is generated through the application layer (such as a map navigation app or in-vehicle navigation system) in the form of maps, text, or voice, and traffic information (such as changes in congested road sections and temporary road closures) is updated in real time, including: • Based on the optimal path node sequence, a visual path is generated in a map navigation app or in-vehicle navigation system through a map rendering engine; • Based on a standardized real-time traffic data interface, the system continuously updates information on changes in congested road sections and temporary road closures through a route refresh mechanism; • Based on user interaction feedback or path deviation warnings, the recommended path is dynamically adjusted through the path recalculation module; • Generate final route recommendations and continuously output them to the user in the form of maps, text, or voice, ensuring that the user is always aware of the real-time status of the travel route.

[0042] It should be noted that the calculated optimal route is presented to the user in an intuitive way, commonly through applications such as map navigation apps and in-vehicle navigation systems. Users input their starting point and destination in these applications, and in this embodiment, after calculating the optimal route based on the user's request, the route is displayed to the user in the form of a map, text, or voice. Simultaneously, traffic information is updated in real time, such as changes in congested road sections and temporary road closures, allowing users to understand the real-time situation of their travel route and make informed decisions.

[0043] Based on the above embodiments, this embodiment will provide a detailed description of step S105: In some embodiments, based on historical transportation data (such as equipment damage rate, delay duration, and actual cost overrun rate), real-time traffic changes (such as temporary congestion, road closures, and severe weather), and feedback from power business systems (such as power grid fault systems and power plant inventory systems), an updated weight strategy is generated and fed back to the multi-objective path selection model through a weight adaptive adjustment mechanism, including: • The weights are dynamically adjusted based on standardized real-time traffic data and emergency data; It should be noted that the model can automatically adjust the weights when "temporary congestion, road closures, or severe weather" occur during transportation. For example, if a sudden traffic jam occurs on the original route during the transportation of a conventional transformer (based on real-time traffic data feedback), and the power plant has no urgent inventory needs, the "timeliness" weight can be automatically reduced (e.g., from 0.2 to 0.1), and the "overall cost" weight can be increased (e.g., from 0.3 to 0.4). The route that is "longer but without congestion and without over-limit fees" can be prioritized to avoid fuel waste caused by congestion.

[0044] • The weights are automatically adjusted based on historical transportation data feedback; It should be noted that the weights are automatically calibrated by accumulating the transportation results of similar tasks (such as "equipment damage rate, delay time, and actual cost overrun rate"). For example, if it is found that a "safety" weight of 0.3 still results in 10% of the equipment being damaged by bumps during multiple transports of precision equipment, the model can automatically increase the "safety" weight to 0.4~0.5, while reducing the "timeliness" weight (e.g., from 0.3 to 0.2), until the damage rate drops to an acceptable range (e.g., ≤2%).

[0045] • Automatically switch weighting strategies based on linkage signals from the power business system; It should be noted that the system connects to the power grid fault system and the power plant inventory system, and automatically switches the weighting strategy according to real-time business needs. For example, when the power grid system issues a "critical fault" signal (such as the main substation being out of service), the model can automatically increase the "timeliness" weight of "emergency repair materials" from 0.6 to 0.7~0.8, and at the same time coordinate with the transportation department to apply for "green wave" to ensure that the materials are delivered with priority. If the power plant's inventory system reports "sufficient coal inventory (50% above the safety threshold)," the "timeliness" weight for coal transportation can be reduced from 0.4 to 0.2, prioritizing low-cost routes.

[0046] • The updated weighting strategy is fed back to the multi-objective path selection model to achieve closed-loop optimization and continuous performance improvement.

[0047] This invention accelerates emergency repairs by dynamically avoiding congestion, shortening the delivery time of power outage repair materials by 25%-40%, reducing power grid restoration time by 15%-30%, and minimizing power outage losses. It also reduces the cost of transporting large items by optimizing compliant routes for special materials such as transformers and wind turbine blades, reducing empty runs by 10%-15%, and lowering unit transportation costs by 8%-12%. Ensuring fuel supply: By linking power plant inventory, fuel supply, and traffic data, transportation bottlenecks can be avoided in advance, ensuring "zero fuel outages" for thermal / gas-fired power plants and reducing fuel consumption; Adapting to new energy scenarios: By integrating charging pile / battery swapping station data, routes with both optimal supply and efficiency can be planned for photovoltaic and wind power equipment, supporting electric transportation and contributing to the "dual-carbon" goal of electricity.

[0048] Please refer to Figure 2 , Figure 2 A structural block diagram of an intelligent dynamic route optimization device based on AI and real-time traffic data provided in this embodiment of the invention; the specific device may include: The data acquisition and preprocessing module 100 is used to generate raw real-time traffic datasets based on sensors, GPS devices and mobile devices distributed in various corners of the city, by real-time monitoring of vehicle speed, traffic flow, driving direction and user travel data, and to generate standardized real-time traffic data through data cleaning, format conversion, standardization and normalization methods. The multi-objective path selection model construction module 200 is used to generate a multi-objective path selection model for power scenarios based on standardized real-time traffic data, power business needs and material characteristics by constructing a comprehensive objective function and weight allocation method. The optimal path node sequence generation module 300 is used to generate the optimal path node sequence based on a multi-objective path selection model and standardized real-time traffic data, through the node comprehensive evaluation function, dynamic actual cost function and dynamic heuristic cost function of the generalized adaptive A* algorithm. The final route recommendation generation module 400 is used to generate a final route recommendation based on the optimal route node sequence, in the form of a map, text, or voice, and update the traffic status in real time. The multi-objective path selection model update module 500 is used to generate an updated weight strategy based on historical transportation data, real-time traffic changes and feedback from the power business system, through a weight adaptive adjustment mechanism, and then feed it back to the multi-objective path selection model.

[0049] The intelligent dynamic route optimization device based on AI and real-time traffic data in this embodiment is used to implement the aforementioned intelligent dynamic route optimization method based on AI and real-time traffic data. Therefore, the specific implementation of the intelligent dynamic route optimization device based on AI and real-time traffic data can be found in the previous embodiment section of the intelligent dynamic route optimization method based on AI and real-time traffic data. For example, the data acquisition and preprocessing module 100, the multi-objective path selection model construction module 200, the optimal path node sequence generation module 300, the final path recommendation generation module 400, and the multi-objective path selection model update module 500 are respectively used to implement steps S101, S102, S103, S104, and S105 in the aforementioned intelligent dynamic route optimization method based on AI and real-time traffic data. Therefore, its specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0050] A specific embodiment of the present invention also provides an intelligent dynamic route optimization device based on AI and real-time traffic data, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the aforementioned intelligent dynamic route optimization method based on AI and real-time traffic data.

[0051] A specific embodiment of the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described intelligent dynamic route optimization method based on AI and real-time traffic data.

[0052] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0053] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0054] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0055] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0056] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for intelligent dynamic route optimization based on AI and real-time traffic data, characterized in that, include: Based on sensors, GPS devices, and mobile devices distributed throughout the city, raw real-time traffic datasets are generated by real-time monitoring of vehicle speed, traffic flow, driving direction, and user travel data. Standardized real-time traffic data is then generated through data cleaning, format conversion, standardization, and normalization methods. Based on standardized real-time traffic data, power business needs and material characteristics, a multi-objective path selection model for power scenarios is generated by constructing a comprehensive objective function and weight allocation method. Based on a multi-objective path selection model and standardized real-time traffic data, the optimal path node sequence is generated through the node comprehensive evaluation function, dynamic actual cost function and dynamic heuristic cost function of the generalized adaptive A* algorithm. Based on the optimal path node sequence, the final path recommendation is generated in the form of map, text or voice, and the traffic status is updated in real time. Based on historical transportation data, real-time traffic changes, and feedback from the power business system, an updated weight strategy is generated through a weight adaptive adjustment mechanism and fed back to the multi-objective path selection model.

2. The intelligent dynamic route optimization method based on AI and real-time traffic data according to claim 1, characterized in that, The method for generating raw real-time traffic datasets based on sensors, GPS devices, and mobile devices distributed throughout the city, through real-time monitoring of vehicle speed, traffic flow, driving direction, and user travel data, includes: Based on geomagnetic sensors, microwave sensors, and cameras, information on vehicle speed, traffic flow, and driving direction on the road is collected; Based on GPS devices, the system collects real-time location and driving trajectory data of vehicles via satellite positioning. Based on applications installed on mobile devices, collect users' travel behavior data; The aforementioned multi-source heterogeneous data are aggregated into the data acquisition layer to generate a raw real-time traffic dataset, which is then transmitted to the data processing layer.

3. The intelligent dynamic route optimization method based on AI and real-time traffic data according to claim 1, characterized in that, The process of generating standardized real-time traffic data through data cleaning, format conversion, standardization, and normalization methods includes: Based on the original real-time traffic dataset, data cleaning algorithms and rules are used to identify and remove noise, errors and duplicate information to generate accurate cleaned data. Based on the cleaned and accurate data, through format conversion processing, the speed data collected by different sensors are unified in units, and the GPS coordinates are converted into a format that the model can process, generating intermediate data with a unified format. Based on the uniformly formatted intermediate data, standardized real-time traffic data that can be used for model processing is generated through standardization and normalization operations. Based on standardized real-time traffic data, statistical analysis and data mining techniques are used to generate traffic flow trends and congested road segment distribution patterns.

4. The intelligent dynamic route optimization method based on AI and real-time traffic data according to claim 1, characterized in that, The multi-objective path selection model for power scenarios, based on standardized real-time traffic data, power business demands, and material characteristics, generates the following by constructing a comprehensive objective function and weight allocation method: Based on the type of power transmission task, initial weights are set for timeliness, overall cost, and transportation safety; Based on the characteristics of power materials, the weight allocation is adjusted to generate a weight combination that is adapted to the characteristics of the materials. Based on the constraints of power business demand, adjust the weights to generate scenario-based weight strategies; By inputting a scenario-based weighting strategy, labeled traffic data, and special electricity costs into a comprehensive objective function, and balancing multiple objectives through a weighted summation method, a multi-objective path selection model is generated.

5. The intelligent dynamic route optimization method based on AI and real-time traffic data according to claim 1, characterized in that, The process of generating the optimal path node sequence based on a multi-objective path selection model and standardized real-time traffic data, using the node comprehensive evaluation function, dynamic actual cost function, and dynamic heuristic cost function of the generalized adaptive A* algorithm, includes: Based on the starting and ending points of power material transportation, the comprehensive evaluation value of each node is calculated through the node comprehensive evaluation function to generate a node priority sequence; Based on standardized real-time traffic data and the special constraints of electric power transportation, the dynamic actual cost from the starting point to each node is calculated through a dynamic actual cost function. Based on real-time traffic trend prediction and power transportation time constraints, the dynamic heuristic cost from each node to the destination is estimated through a dynamic heuristic cost function. By comprehensively evaluating all nodes, the optimal path node sequence is generated through a generalized adaptive A* algorithm search.

6. The intelligent dynamic route optimization method based on AI and real-time traffic data according to claim 1, characterized in that, The process of generating a final route recommendation based on the optimal path node sequence, in the form of a map, text, or voice, and updating traffic status in real time includes: Based on the optimal path node sequence, a visual path is generated in a map navigation APP or in-vehicle navigation system through a map rendering engine. Based on a standardized real-time traffic data interface, the route refresh mechanism continuously updates information on changes in congested road sections and temporary road control measures. Based on user interaction feedback or path deviation warnings, the recommended path is dynamically adjusted through the path recalculation module. Generate the final route recommendation and continuously output it to the user in the form of map, text or voice.

7. The intelligent dynamic route optimization method based on AI and real-time traffic data according to claim 1, characterized in that, The process of generating an updated weight strategy based on historical transportation data, real-time traffic changes, and feedback from the power business system, and feeding it back to the multi-objective path selection model through a weight adaptive adjustment mechanism, includes: The weights are dynamically adjusted based on standardized real-time traffic data and emergency data. The weights are automatically adjusted based on historical transportation data feedback. Automatically switch weighting strategies based on linkage signals from the power business system; The updated weighting strategy is fed back into the multi-objective path selection model.

8. An intelligent dynamic route optimization device based on AI and real-time traffic data, characterized in that, include: The data acquisition and preprocessing module is used to generate raw real-time traffic datasets based on sensors, GPS devices and mobile devices distributed throughout the city, by real-time monitoring of vehicle speed, traffic flow, driving direction and user travel data. It also generates standardized real-time traffic data through data cleaning, format conversion, standardization and normalization methods. The multi-objective path selection model construction module is used to generate a multi-objective path selection model for power scenarios based on standardized real-time traffic data, power business needs and material characteristics by constructing a comprehensive objective function and weight allocation method. The optimal path node sequence generation module is used to generate the optimal path node sequence based on a multi-objective path selection model and standardized real-time traffic data, through the node comprehensive evaluation function, dynamic actual cost function and dynamic heuristic cost function of the generalized adaptive A* algorithm. The final route recommendation generation module is used to generate final route recommendations based on the optimal route node sequence, in the form of maps, text, or voice, and update traffic status in real time. The multi-objective path selection model update module is used to generate an updated weight strategy based on historical transportation data, real-time traffic changes, and feedback from the power business system, and then feed it back to the multi-objective path selection model through a weight adaptive adjustment mechanism.

9. An intelligent dynamic route optimization device based on AI and real-time traffic data, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the intelligent dynamic route optimization method based on AI and real-time traffic data as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the intelligent dynamic route optimization method based on AI and real-time traffic data as described in any one of claims 1 to 7.

Images