A logistics transportation rapid planning method based on big data analysis

Abstract

The application discloses a logistics transportation rapid planning method based on big data analysis, and relates to the technical field of intelligent logistics and traffic planning, and comprises the following steps: acquiring traffic event data, road network data, position data and cold-chain cargo state data; performing nonlinear weighted calculation on the traffic event data, and outputting influence coefficients of the traffic event data on each road section; and obtaining the passable scores of each road section through neural network output according to the influence coefficients and the road network data of the corresponding road sections. The scheme combines dynamic threshold values and path integrity checking, avoids the path unreachability problem caused by rigid threshold values, adjusts and selects conditions according to cargo characteristics, for example, the threshold value of temperature-sensitive cargo is increased to preferentially select road sections with higher passing efficiency, solves the problems of insufficient path reliability and poor cargo adaptability caused by fixed threshold values in the prior art, and ensures that the generated initial bypassable road sections meet the preservation requirements of the cargo.

Description

Technical Field

[0001] This invention relates to the field of intelligent logistics and transportation planning technology, specifically a rapid logistics transportation planning method based on big data analysis. Background Technology

[0002] Modern logistics and transportation systems are facing increasingly complex road network environments and challenges from sudden traffic incidents. In particular, in the field of cold chain cargo transportation, with the development of intelligent transportation systems and big data technology, dynamic route planning based on real-time data analysis has become a key technology for improving logistics efficiency. Cold chain transportation has strict requirements for timeliness, temperature stability and route reliability.

[0003] However, in the process of implementing the technical solution of the invention in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0004] Existing technologies cannot effectively handle the combined impact of various traffic events, such as traffic accidents, road construction, and extreme weather, on cold chain transportation routes. This results in significant discrepancies between planning results and actual road conditions. When selecting alternative routes, only road capacity is considered without adaptive filtering based on the real-time status of cold chain goods. These problems severely restrict the transportation efficiency and cargo safety of cold chain logistics. Summary of the Invention

[0005] The purpose of this invention is to provide a rapid logistics transportation planning method based on big data analysis to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] In a first aspect, this invention discloses a rapid logistics transportation planning method based on big data analysis, applied to adaptive planning of cold chain cargo transportation when transportation routes are temporarily closed, comprising the following steps:

[0008] Acquire traffic incident data, road network data, location data, and cold chain cargo status data;

[0009] The traffic incident data is subjected to nonlinear weighted calculation to output the impact coefficient of the traffic incident data on each road segment;

[0010] Based on the influence coefficient and the road network data of the corresponding road segment, the passability score of each road segment is obtained through the output of a neural network.

[0011] Determine whether the passability score is greater than a preset threshold; otherwise, remove the road segment and obtain the initial detourable road segments.

[0012] Based on the road network data, the predicted values ​​of traffic flow and carrying capacity of each initial detour route are obtained through time-series prediction.

[0013] Determine whether the predicted value of the carrying capacity data is greater than the set value. If so, remove the initial detourable road segment and obtain the detourable road segment.

[0014] Based on the predicted traffic flow and carrying capacity data of the detour routes and the corresponding routes, the estimated travel time of each detour route is calculated. Detour routes with inconsistent estimated travel times are then eliminated by combining the cold chain cargo status data, and the final detour routes are obtained.

[0015] Based on the location data, the final detour route, and the estimated travel time, a route is planned and a cold chain transportation route is output.

[0016] Secondly, this invention discloses a rapid logistics transportation planning system based on big data analysis, comprising:

[0017] The data acquisition module is used to acquire traffic incident data, road network data, location data, and cold chain cargo status data.

[0018] The passability score calculation module is used to perform nonlinear weighted calculation on the traffic incident data and output the impact coefficient of the traffic incident data on each road segment.

[0019] Based on the influence coefficient and the road network data of the corresponding road segment, the passability score of each road segment is obtained through the output of a neural network.

[0020] The initial detourable road segment determination module is used to determine whether the passability score is greater than a preset threshold; otherwise, the road segment is removed to obtain the initial detourable road segment.

[0021] The detourable road segment identification module is used to obtain the predicted values ​​of traffic flow data and carrying capacity data of each initial detourable road segment through time-series prediction based on the road network data.

[0022] Determine whether the predicted value of the carrying capacity data is greater than the set value. If so, remove the initial detourable road segment and obtain the detourable road segment.

[0023] The final detour route output module calculates the estimated travel time of each detour route based on the detour routes and the predicted values ​​of traffic flow and carrying capacity data of the corresponding routes. It also eliminates detour routes with inconsistent predicted travel times by combining the cold chain cargo status data, thus obtaining the final detour routes.

[0024] The cold chain transportation route planning module is used to plan routes and output cold chain transportation routes based on the location data, the final detour route, and the estimated travel time.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] 1. This solution combines dynamic thresholds with path integrity verification, which avoids the path inaccessibility problem caused by rigid thresholds and adjusts the screening conditions according to the characteristics of the goods. For example, the threshold is increased for temperature-sensitive goods to prioritize the selection of road segments with higher traffic efficiency. This solves the problems of insufficient path reliability and poor adaptability of goods caused by fixed thresholds in the existing technology, ensuring that the generated initial detour routes meet the freshness requirements of the goods and have complete accessibility.

[0027] 2. This solution, by simultaneously predicting traffic flow and carrying capacity, makes the travel time calculation closer to the actual road conditions, avoiding the selection of routes that cannot actually meet the needs of cold chain transportation due to ignoring carrying capacity limitations, and can accurately assess the actual traffic efficiency of alternative routes.

[0028] 3. This solution transforms the temperature control requirements of cold chain transportation into a quantitative indicator for route selection by introducing the remaining shelf life. This solves the problem of route failure caused by ignoring the condition of goods in traditional methods, and achieves precise matching between cold chain transportation routes and the shelf life of goods, effectively avoiding the risk of goods spoilage caused by excessive detour time. Attached Figure Description

[0029] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0030] Figure 1 This is a flowchart illustrating the steps of a rapid logistics transportation planning method based on big data analysis according to the present invention.

[0031] Figure 2 A schematic diagram illustrating the process of obtaining the initial detourable road segment provided by the present invention;

[0032] Figure 3 A flowchart illustrating the process of calculating the estimated travel time for each detour route provided by this invention;

[0033] Figure 4 This invention provides a schematic diagram of the module functions of a logistics transportation rapid planning system based on big data analysis. Detailed Implementation

[0034] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0035] Application Overview:

[0036] In existing technologies, cold chain cargo transportation faces challenges from complex road network environments and sudden traffic events. Existing dynamic route planning methods struggle to effectively handle the combined impacts of various traffic events, such as the reduced road capacity caused by the combined effects of accidents, construction, and weather. Traditional methods, when selecting alternative routes, do not incorporate status data such as the remaining shelf life of cold chain cargo, which can easily lead to the selection of routes with travel times exceeding the cargo's shelf life, resulting in cargo damage. Furthermore, existing time-series forecasting models only focus on traffic flow data and do not simultaneously analyze changes in the carrying capacity of alternative routes, leading to insufficient reliability in route planning.

[0037] To address the aforementioned issues and the need for dynamic assessment of the combined impact of multiple events, a nonlinear weighted model is used to quantify the cumulative impact of different types of traffic events on road segments. To address the correlation between the status of cold chain goods and route planning, a dynamic matching of goods shelf life and road segment travel time is proposed. To resolve the coordination issue between carrying capacity and traffic flow prediction, a prediction model integrating temporal features is constructed to synchronously output traffic flow and carrying capacity data. Finally, complex road network data is processed through neural networks, combined with a multi-dimensional filtering mechanism to achieve accurate route planning.

[0038] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0039] Example 1:

[0040] Please see Figure 1 - Figure 3 A rapid logistics transportation planning method based on big data analysis is applied to adaptive planning of cold chain cargo transportation when transportation routes are temporarily closed. The method includes the following steps:

[0041] Acquire traffic incident data, road network data, location data, and cold chain cargo status data;

[0042] Perform nonlinear weighted calculations on traffic incident data and output the impact coefficients of traffic incident data on each road segment;

[0043] Based on the influence coefficient and the road network data of the corresponding road segment, the passability score of each road segment is obtained through the output of a neural network;

[0044] Determine if the passability score is greater than a preset threshold; otherwise, remove the road segment and obtain the initial detourable road segments.

[0045] Based on road network data, time-series predictions are used to obtain the predicted traffic flow and carrying capacity data for each initial detour route.

[0046] Determine whether the predicted value of the carrying capacity data is greater than the set value. If so, remove the initial detourable road segment and obtain the detourable road segments.

[0047] Based on the predicted traffic flow and carrying capacity data of the detour routes, the estimated travel time of each detour route is calculated. Detour routes with inconsistent estimated travel times are then eliminated by combining the cold chain cargo status data, and the final detour routes are obtained.

[0048] Based on location data, the final detour route, and the estimated travel time, route planning is performed and a cold chain transportation route is output.

[0049] Traffic incident data includes the type of incident, the road segment where it occurred, and the duration, which are obtained from satellite remote sensing platforms and used to dynamically assess the traffic status of the road segment.

[0050] The nonlinear weighted calculation adopts an adaptive weight adjustment mechanism, which dynamically allocates weight coefficients according to the actual impact of different event types on traffic flow.

[0051] The neural network adopts a graph neural network structure, which aggregates traffic features of adjacent road segments through graph convolutional layers and combines an attention mechanism to enhance the ability to extract key features;

[0052] The passability score is processed by normalization and output by the Sigmoid function to quantify the real-time passability of road segments;

[0053] Time series forecasting uses a sliding window method to construct a supervised learning dataset, simultaneously predicting the changing trends of traffic and carrying capacity;

[0054] The estimated travel time calculation takes into account the dynamic relationship between road segment length, traffic flow, and carrying capacity to ensure the accuracy of the time estimate.

[0055] The specific implementation process is as follows:

[0056] First, real-time traffic event data is collected, including information on road closures, traffic accidents, and road construction, and then processed in conjunction with road network data, location data, and cold chain cargo status data. Based on different traffic events, a non-linear weighted calculation method is used to analyze the impact coefficient of each traffic event on different road segments.

[0057] Then, by combining the impact coefficient with the road network data of the corresponding road segment, a neural network model is used to score each road segment and output the passability score of each road segment.

[0058] Determine whether the passability score of each road segment is greater than a preset threshold. If the condition is not met, the road segment is removed.

[0059] Based on this, the traffic flow and carrying capacity data of the initial detourable road segments are predicted through time series prediction. The traffic flow and carrying capacity of each road segment are predicted within a certain time range in the future. If the predicted carrying capacity data of a certain road segment is less than the set value, the road segment will be eliminated to ensure that the selected route can safely carry the transportation needs of cold chain goods.

[0060] Then, by combining the predicted traffic flow data and carrying capacity data, the estimated travel time of each detour route is calculated. For cold chain goods, the estimated travel time is not only related to traffic flow and route length, but also closely related to the temperature, humidity, and condition data of the cold chain goods. For example, fresh products may be more sensitive to transportation temperature, while some pharmaceutical products have higher time requirements. By comprehensively analyzing these data, routes with estimated travel times that do not meet the requirements for cold chain goods transportation are eliminated, thus obtaining the final detour routes.

[0061] Based on the final selected detour routes, route planning is performed using location data, such as using Dijkstra's algorithm to calculate the optimal route from the starting point to the destination, ensuring that cold chain goods can be transported safely in the shortest possible time.

[0062] Compared with existing technologies, this application can effectively handle the combined impact of multiple types of traffic events and accurately assess the real-time traffic capacity of alternative road sections; by synchronously predicting traffic flow and carrying capacity data, it avoids selecting alternative routes that are about to reach their carrying capacity limits; by dynamically selecting road sections based on the remaining shelf life of cold chain goods, it ensures that the passage time of the transportation route is strictly controlled within the shelf life of the goods; and by using graph neural networks to process complex road network features, it improves the calculation accuracy and efficiency of the passability score, ultimately forming a cold chain transportation route planning scheme that takes into account timeliness, reliability, and temperature control stability.

[0063] This application further proposes to obtain traffic incident data, road network data, location data, and cold chain cargo status data, specifically including:

[0064] By connecting with the satellite remote sensing data platform, traffic incident data can be obtained in real time. The traffic incident data includes the type of incident, the road section where the incident occurred, the time of occurrence, the scope of impact of the incident, and the duration of the incident.

[0065] Road network data is obtained from the map service platform. The road network data includes information such as the geographical location, length, real-time traffic data, and maximum carrying capacity of each road segment.

[0066] Location data is acquired in real time through the transportation management system. The location data includes information such as the vehicle's current geographic coordinates, destination coordinates, and transportation trajectory.

[0067] The vehicle's IoT monitoring system acquires cold chain cargo status data, which includes information such as cargo type, current temperature, and remaining shelf life.

[0068] Among them, the satellite remote sensing data platform refers to the system platform that monitors ground traffic events in real time through satellite remote sensing technology. Specifically, it can be achieved by fusing satellite image processing with ground sensor data to capture sudden traffic events such as road closures, accidents, or abnormal weather.

[0069] A map service platform refers to a third-party data service interface that provides real-time road network information. This can be implemented using open API calls to obtain static attributes and dynamic traffic information for road segments. A transportation management system refers to a software system used to manage vehicle transportation tasks. This can be implemented using GPS positioning modules and cloud data synchronization technology to track vehicle locations and transportation trajectories in real time.

[0070] The Internet of Things (IoT) monitoring system refers to a sensor network deployed in cold chain transport vehicles. Specifically, it can be implemented by combining temperature sensors, timers, and wireless transmission modules to monitor the temperature status of goods and the remaining shelf life.

[0071] The satellite remote sensing data platform receives satellite images and ground sensor data to identify the location and impact range of traffic events in real time. For example, when a traffic accident causes the closure of a road section, the platform will push the event type, time of occurrence and expected duration to the route planning system.

[0072] The map service platform obtains the road network topology of the target area by calling the API interface. For example, if the maximum carrying capacity of a certain road segment is 500 vehicles per hour and the real-time traffic flow is 300 vehicles, the system can assess the traffic potential of the road segment based on this.

[0073] The transportation management system collects the latitude and longitude coordinates of vehicles through a GPS module. For example, if the vehicle's current location is 30 degrees north latitude and 120 degrees east longitude, and its destination is 31 degrees north latitude and 121 degrees east longitude, the system can generate transportation trajectory data.

[0074] The IoT monitoring system collects temperature data inside the vehicle compartment through temperature sensors. For example, if the current temperature is -18℃ and the remaining freshness time is 8 hours, the system can use this data to filter out paths that meet the freshness time constraints.

[0075] Beneficial effects:

[0076] This application can integrate multi-source heterogeneous data in real time. For example, when cold chain transport vehicles encounter road closures, the system can quickly identify alternative routes based on satellite remote sensing data, exclude overloaded routes by combining road network carrying capacity data provided by the map service platform, and eliminate routes with excessively long estimated travel times based on the remaining shelf life fed back by the Internet of Things monitoring system, thereby ensuring that cold chain goods arrive at their destination within the shelf life.

[0077] This application further proposes to perform nonlinear weighted calculation on traffic incident data, and the output impact coefficients of traffic incident data on each road segment specifically include:

[0078] Multi-dimensional feature extraction is performed on the collected traffic incident data;

[0079] Multi-dimensional features are processed through a non-linear weighted model to output the initial impact coefficients of traffic event data on each road segment; the non-linear weighted model can automatically adjust the weight coefficients of each feature according to the actual impact of different event types on traffic flow and road capacity.

[0080] The initial impact coefficient of each traffic incident data on each road segment is assigned to the corresponding road segment;

[0081] For the same road segment affected by multiple traffic incident data, the initial impact coefficient of each traffic incident data is weighted and calculated to obtain the final impact coefficient of the road segment.

[0082] Multi-dimensional feature extraction refers to extracting feature parameters of different dimensions such as event type, duration, and scope of impact from traffic event data. Specifically, feature engineering methods combined with domain knowledge can be used for screening, such as dimensionality reduction through principal component analysis. This step can comprehensively capture the multifaceted impact of traffic events on the road network, providing a data foundation for subsequent weight adjustments.

[0083] Among them, the nonlinear weighted model refers to a mathematical model that dynamically assigns weights to different features through a nonlinear function. Specifically, it can be implemented using a multilayer perceptron or a support vector machine. This model can adaptively adjust the importance of each feature based on the nonlinear relationship between event type and traffic flow changes, such as the exponential decay effect of traffic accidents on traffic capacity, thus avoiding the accumulation of errors caused by linear weighting.

[0084] Weighted calculation refers to a comprehensive evaluation method for the same road segment affected by multiple events. Specifically, it can adopt weighted average or weighted summation algorithms, such as setting a time decay factor based on the time of the events. This step can effectively integrate the superimposed effects of multiple events on the same road segment and accurately reflect the comprehensive impact of composite events on the road segment.

[0085] The specific implementation steps are as follows: Multi-dimensional feature extraction is performed on the collected traffic incident data: For each traffic incident data, features such as the type of the incident, the time of occurrence, the duration, and the location are extracted, as well as the potential influencing factors of these incidents, such as the scale of the accident, weather conditions, and the length of the construction area.

[0086] A nonlinear weighted model is used to process traffic event data of different types. This model can dynamically adjust the weight coefficients of each feature according to the actual impact of different events on traffic flow and road capacity. Specifically, the model adjusts the weights based on the following factors:

[0087] Event type: Events such as traffic accidents, road construction, and weather changes have different impacts on traffic flow. The model will automatically learn the weight value of each type of event based on historical data.

[0088] Event scale: Larger-scale accidents or construction projects have a greater impact on traffic, and the model will automatically increase the weight of such events.

[0089] Time factor: The occurrence of events during peak and off-peak hours affects the degree of impact on traffic flow; by adjusting the weights according to time period, the model can more accurately reflect the impact of traffic events.

[0090] By weighting these features, the initial impact coefficient of each traffic event data on each road segment is output.

[0091] For each road segment, the initial impact coefficient of the road segment is calculated by weighting based on traffic event data. In particular, for road segments affected by multiple traffic events at the same time, the impact coefficients of these events are combined and weighted to obtain the final road segment impact coefficient.

[0092] Through the above technical solution, this application can effectively solve the problem of assessing the combined impact of multiple types of traffic events on cold chain transportation routes; through a dynamic weight adjustment mechanism, it can accurately quantify the differentiated impact of different types of events on the traffic capacity of road sections, avoiding route selection errors caused by unreasonable weight settings. For example, when temporary traffic control and short-term heavy rainfall occur simultaneously, this solution can accurately distinguish the differences in the impact of the two types of events on the timeliness of cold chain vehicle passage, providing reliable data support for subsequent route planning.

[0093] This application further proposes a method to obtain the passability score of each road segment based on the influence coefficient and the road network data of the corresponding road segment through the output of a neural network, specifically including:

[0094] The influence coefficients and the corresponding road network data are normalized.

[0095] A neural network is constructed, specifically a graph neural network. Its topology is built based on actual road network data, where nodes represent intersections and edges represent road segments. Traffic impact features of adjacent road segments are aggregated through graph convolutional layers, specifically including:

[0096] The input layer receives the input of influence coefficients and road network data;

[0097] At least one hidden layer employs the ReLU activation function for nonlinear transformation and introduces an attention mechanism to dynamically calculate the weights of each input feature;

[0098] The output layer uses the Sigmoid function to output the passability score of the road segment;

[0099] Historical traffic event data, historical road network data, and the passability scores of corresponding road segments are used as supervisory signals to train the neural network using supervised learning.

[0100] The normalized influence coefficients and the corresponding road network data are input into the trained neural network to output the passability score of each road segment.

[0101] Normalization refers to converting influence coefficients and road network data with different dimensions into data with a unified dimension. Specifically, it can be achieved by using maximum-minimum normalization or standardization methods to ensure that the numerical range of the neural network input is consistent.

[0102] Among them, graph neural networks refer to deep learning models based on graph structure data. Their topology corresponds to the actual road network. Specifically, they can be implemented using graph convolutional networks, which capture the spatial correlation between road segments by aggregating the feature information of adjacent nodes.

[0103] The attention mechanism refers to the calculation module that dynamically adjusts the weights of input features. Specifically, it can be implemented using a self-attention mechanism or a multi-head attention mechanism, which automatically focuses on key features based on the current traffic events and road network status.

[0104] Supervised learning refers to using real passability scores from historical data as training targets. Specifically, backpropagation algorithms can be used to optimize neural network parameters and improve model accuracy by minimizing the difference between predicted and real scores.

[0105] The specific process is as follows:

[0106] The influence coefficient and the corresponding road network data are normalized to ensure that all data are within the same range, thus avoiding the excessive dominance of certain features in model training.

[0107] By constructing a graph neural network model, traffic impact information of adjacent road segments can be effectively aggregated;

[0108] In the input layer of the network, the influence coefficients and road network data are input into the neural network. Then, the hidden layer of the network uses the ReLU activation function for nonlinear transformation and introduces an attention mechanism to dynamically adjust the weights of the input features according to the different importance of the data features during training. Through the graph convolutional layer, the traffic influence features of adjacent road segments of each road segment are aggregated, so that the model can effectively understand the complex relationships between road networks.

[0109] The neural network is trained using supervised learning. The training data includes historical traffic event data, road network data, and the accessibility scores of corresponding road segments. The network weights are updated by calculating the error between the actual accessibility score and the predicted score to improve the prediction accuracy.

[0110] After training, the normalized traffic event data and road network data are input into the trained neural network to obtain the passability score for each road segment.

[0111] Beneficial effects:

[0112] This application solves the problem of inaccurate scoring caused by ignoring the road network topology and dynamic weights of features in the prior art, and improves the reliability and real-time performance of passability scoring. For example, when cold chain transport vehicles encounter temporary closures of multiple road sections, this solution can quickly identify alternative routes with sufficient capacity and that meet the requirements for the freshness of goods, avoiding the selection of road sections with excessively long travel times due to scoring errors, thereby ensuring the quality stability of cold chain goods.

[0113] This application further proposes a method to determine whether the passability score is greater than a preset threshold, and if not, to remove the road segment, thus obtaining the initial detour road segments, specifically including:

[0114] Determine if the passability score is greater than a preset threshold; otherwise, remove the road segment and obtain the initial detourable road segments.

[0115] Verify whether the initial detour route meets the requirement of a complete path from the vehicle's current geographical coordinates to the destination coordinates based on the location data;

[0116] When the verification result is not met, the initial detour routes are supplemented according to the passability score from high to low until the initial detour routes satisfy the complete path from the vehicle's current geographical coordinates to the destination coordinates.

[0117] The preset threshold is a dynamic variable related to the status data of cold chain goods, and its specific calculation formula is as follows:

[0118]

[0119] In the formula, Indicates the preset threshold. Indicates the basic threshold. This indicates that the threshold is adjusted based on the data of cold chain goods types, for temperature-sensitive cold chain goods (such as vaccines). For regular cold chain goods (such as popsicles). For heat-resistant cold chain goods (such as certain medicines). .

[0120] The preset threshold refers to the minimum scoring standard used to screen detour routes. It can be implemented by dynamic calculation. The basic threshold is adjusted by combining the type of cold chain goods. For example, temperature-sensitive goods require higher passage reliability, thereby improving the screening standard.

[0121] Dynamic variables refer to preset thresholds that change dynamically according to the preservation requirements of goods. Specifically, cold chain goods type data can be used as input parameters. For example, vaccines correspond to higher adjustment thresholds to ensure that route selection prioritizes time sensitivity.

[0122] Cold chain cargo type data refers to the classification information of temperature control requirements of goods during transportation. Specifically, it can be collected by the Internet of Things monitoring system and transmitted to the planning system. For example, goods can be divided into temperature-sensitive, conventional, and temperature-resistant types, each corresponding to different threshold adjustment strategies.

[0123] The implementation process is as follows:

[0124] The system judges based on the passability score and the preset threshold. If the passability score of a certain road segment is lower than the threshold, it is removed to obtain the initial detour road segments. The preset threshold is a dynamic variable that can be adjusted according to the changes in the status of cold chain goods.

[0125] Verify the initial detour routes based on location data: By comparing the coordinates of the starting point and the ending point, verify whether the initial detour routes can form a complete path based on the current location. If the path is incomplete, i.e., it is impossible to reach the ending point smoothly from the starting point, supplement more detour routes according to the passability score from high to low, until the requirement of a complete path from the starting point to the ending point is met.

[0126] Through the above technical solution, this application solves the problems of insufficient path reliability and poor cargo adaptability caused by fixed thresholds in the prior art. By dynamically adjusting the screening criteria and combining path topology verification, it ensures that the generated initial detourable road segments not only meet the cargo preservation requirements but also have complete accessibility. For example, when transporting vaccines, the system automatically increases the threshold and supplements high-scoring road segments to avoid path interruption due to a single event, while ensuring that the transportation time is within the remaining preservation time.

[0127] This application further proposes to obtain the predicted traffic flow data and carrying capacity data of each initial detour segment through time-series prediction based on road network data, specifically including:

[0128] Construct a prediction model that integrates temporal features;

[0129] Extract historical traffic flow data and historical maximum carrying capacity data of road segments from historical road network data;

[0130] Historical traffic data and historical maximum carrying capacity data of road segments were converted into supervised learning datasets using the sliding window method.

[0131] The prediction model is trained using a supervised learning dataset;

[0132] The road network data is input into the prediction model, and the predicted values ​​of traffic flow and carrying capacity data for each initial detour segment are output.

[0133] Among them, the prediction model that integrates time series features refers to a machine learning model that can capture the dynamic change patterns in time series data. Specifically, it can be implemented using long short-term memory networks or autoregressive integral moving average models, and is used to simultaneously predict the future changing trends of traffic and carrying capacity.

[0134] The sliding window method refers to dividing time series data into multiple training samples according to a fixed window length. Specifically, a sliding method with a window length of 5-10 time steps can be used to convert historical traffic and carrying capacity data into input-output pairs to meet the training requirements of supervised learning models.

[0135] Supervised learning datasets refer to datasets containing input features and target labels. Specifically, the first 6 hours of historical traffic data can be used as input features, and the last 2 hours as target labels, to train models to predict future traffic and carrying capacity.

[0136] Specifically, the sliding window method is used to divide historical traffic data and maximum capacity data into multiple time windows. Each window contains historical data over a period of time. By using the sliding window method, the data in each time window is paired with the corresponding label (i.e., the future values ​​of traffic and capacity) to form a supervised learning dataset. This not only captures the temporal characteristics of traffic and capacity, but also predicts future traffic changes through historical data.

[0137] The supervised learning algorithm is used to train the supervised learning dataset. During the training process, cross-validation is used to avoid overfitting, and the prediction effect is optimized by adjusting the model parameters.

[0138] After the model training is completed, the current road network data is input into the trained prediction model, and the predicted values ​​of traffic flow and carrying capacity data for each initial detour segment are output.

[0139] Through the above technical solution, this application solves the problem that the time series prediction model in the prior art only focuses on traffic flow data and ignores changes in carrying capacity. It realizes a dual assessment of the traffic capacity of detour sections, ensuring that the selected sections meet the cold chain transportation requirements in both traffic flow and carrying capacity dimensions, thereby reducing the risk of transportation interruption caused by road section overload and improving the effectiveness of route planning.

[0140] This application further proposes to calculate the estimated travel time for each detour route based on the detour route and the corresponding traffic flow data prediction values, specifically including:

[0141] The length of the detour route is obtained from the road network data corresponding to the detour route.

[0142] Based on the predicted traffic flow and carrying capacity data of the alternative route, the traffic velocity of the alternative route is calculated using the following formula:

[0143]

[0144] In the formula, Indicates alternative routes In time Traffic flow rate This represents the proportionality coefficient. Indicates alternative routes In time The predicted value of carrying capacity data, Indicates alternative routes In time Traffic data prediction values;

[0145] The estimated travel time for the alternative route is calculated based on the traffic flow speed and length of the alternative route. The specific calculation formula is as follows:

[0146]

[0147] In the formula, Indicates alternative routes In time The estimated travel time, Indicates alternative routes The length.

[0148] Among them, the traffic flow data prediction value refers to the value obtained by predicting the traffic flow of detour road sections at future time points through a time series prediction model. Specifically, it can be implemented by using the sliding window method combined with the long short-term memory network model, and is used to reflect the future traffic congestion trend of the road section.

[0149] Among them, the carrying capacity data prediction value refers to the value obtained by predicting the maximum traffic flow that can be accommodated at future time points of detour road sections through a time series prediction model. Specifically, it can be implemented using the support vector regression algorithm to evaluate the traffic limit of the road section within the predicted time.

[0150] Specifically, the length of detour routes is determined using road network data; traffic flow speed is calculated based on the predicted traffic volume and carrying capacity of each detour route; and the estimated travel time for each detour route is calculated based on its traffic flow speed and length.

[0151] Beneficial effects:

[0152] This application can accurately assess the actual traffic efficiency of alternative routes and dynamically eliminate overdue routes based on the remaining shelf life of cold chain goods. This provides a detour solution for cold chain transportation that balances timeliness and reliability in the event of a sudden traffic incident, effectively reducing the risk of spoilage of goods due to transportation delays.

[0153] This application further proposes to eliminate alternative routes that do not match the estimated travel time by combining cold chain cargo status data, resulting in the following final alternative routes:

[0154] The minimum remaining shelf life is calculated from the status data of cold chain goods.

[0155] Determine whether the estimated travel time is greater than the minimum remaining shelf life. If so, remove the detour route and obtain the final detour route.

[0156] The remaining shelf life refers to the maximum time threshold under current temperature conditions for cold chain goods to maintain quality and safety. Specifically, it can be calculated by collecting real-time temperature data of goods through IoT sensors and combining it with a preset temperature decay model for each type of goods.

[0157] Specifically, the system iterates through the estimated travel time data for all possible detour routes and extracts the remaining shelf life information from the cold chain cargo status data. By comparing the estimated travel time of each route with the minimum remaining shelf life, routes with travel times exceeding the shelf life constraint are automatically filtered out. For example, if the estimated travel time of a detour route is 5 hours, and the minimum remaining shelf life of the cargo is 4 hours, the route will be removed by the system. This ensures that the final output of detour routes meets the timeliness requirements for cold chain cargo transportation.

[0158] Through the above technical solution, this application achieves precise matching between cold chain transportation routes and the shelf life of goods, effectively avoiding the risk of goods spoilage caused by excessive detour time; by dynamically eliminating route options that do not meet the shelf life constraints, it ensures that the transportation plan can still meet the reliability requirements of cold chain logistics in the event of a sudden traffic incident.

[0159] This application further proposes to plan routes and output cold chain transportation routes based on location data, final detour routes, and estimated travel time, specifically including:

[0160] Obtain the vehicle's current geographic coordinates and destination coordinates;

[0161] The final detour route will be used as input data for route planning;

[0162] The shortest path algorithm is used to calculate the optimal path between the vehicle's current geographical coordinates and the destination coordinates, based on the estimated travel time of each road segment. The optimal path is the path with the lowest estimated travel time, and the estimated travel time is less than the minimum of the remaining shelf life.

[0163] Output the optimal path and perform dynamic tracking.

[0164] Location data refers to the vehicle's current geographical coordinates, destination coordinates, and transportation trajectory, which are obtained in real time through the transportation management system. Specifically, it can be implemented using a GPS positioning module or the BeiDou navigation system to determine the starting and ending points of the route planning.

[0165] The final detour route refers to the alternative route that meets the traffic conditions and the shelf life limit of cold chain goods after multiple screenings. Specifically, it can be achieved by using a multi-stage filtering algorithm that combines the traffic event impact coefficient, the carrying capacity prediction value, and the remaining shelf life to exclude infeasible routes.

[0166] Dynamic tracking refers to real-time monitoring and adjustment of the output optimal path. Specifically, it can be achieved through a data interaction mechanism between in-vehicle IoT devices and cloud-based path planning systems to cope with sudden traffic events or road condition fluctuations.

[0167] The specific process is as follows: First, the real-time location information of the vehicle and the coordinates of the destination are obtained. The final detour route segment after multi-stage screening is used as a candidate route set. An improved Dijkstra algorithm is used, with the estimated travel time as a weight factor, to search for the shortest time path from the origin to the destination in the candidate route set. During the execution of the algorithm, the remaining shelf life is set as a constraint condition to automatically exclude routes whose estimated travel time exceeds the shelf life of the goods. After the route is generated, the vehicle terminal and the cloud server communicate continuously to monitor changes in road conditions in real time and trigger route replanning to ensure the timeliness and reliability of cold chain transportation.

[0168] In some specific implementations, the shortest path algorithm can be optimized using AI algorithms, such as introducing heuristic functions to estimate the travel time of the remaining road segments and improve search efficiency; the dynamic tracking function can be implemented by setting a periodic location reporting mechanism, such as sending vehicle coordinates to the cloud every 5 minutes, and the server determines whether to trigger a path update.

[0169] Beneficial effects:

[0170] This application can effectively avoid the problem of cargo damage caused by the route travel time exceeding the shelf life of cold chain goods. It improves the reliability and adaptability of transportation routes through dynamic optimization and real-time monitoring, and solves the problem of route planning being disconnected from cargo status in the prior art.

[0171] Example 2:

[0172] Please see Figure 4 A rapid logistics transportation planning system based on big data analysis includes:

[0173] The data acquisition module is used to acquire traffic incident data, road network data, location data, and cold chain cargo status data.

[0174] The passability score calculation module is used to perform non-linear weighted calculations on traffic incident data and output the impact coefficient of traffic incident data on each road segment.

[0175] Based on the influence coefficient and the road network data of the corresponding road segment, the passability score of each road segment is obtained through the output of a neural network;

[0176] The initial detourable road segment judgment module is used to determine whether the passability score is greater than a preset threshold. Otherwise, the road segment is removed to obtain the initial detourable road segments.

[0177] The detour route identification module is used to obtain the predicted values ​​of traffic flow data and carrying capacity data of each initial detour route through time-series prediction based on road network data;

[0178] Determine whether the predicted value of the carrying capacity data is greater than the set value. If so, remove the initial detourable road segment and obtain the detourable road segments.

[0179] The final detour route output module calculates the estimated travel time for each detour route based on the detour routes and the predicted traffic flow and carrying capacity data of the corresponding routes. It also eliminates detour routes with inconsistent predicted travel times by combining the cold chain cargo status data, thus obtaining the final detour routes.

[0180] The cold chain transportation route planning module is used to plan routes and output cold chain transportation routes based on location data, final detour routes, and estimated travel time.

[0181] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0182] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A logistics transportation rapid planning method based on big data analysis, applied to the adaptive planning of cold chain cargo transportation in the case of temporary closure of the transportation section, characterized in that, Includes the following steps: Acquire traffic incident data, road network data, location data, and cold chain cargo status data; among which, cold chain cargo status data is acquired through the vehicle's IoT monitoring system, including cargo type, current temperature, and remaining shelf life; The Internet of Things (IoT) monitoring system refers to a sensor network deployed in cold chain transport vehicles. It is implemented by combining temperature sensors, timers, and wireless transmission modules to monitor the temperature status of goods and the remaining shelf life. The traffic incident data is subjected to nonlinear weighted calculation to output the impact coefficient of the traffic incident data on each road segment; Based on the influence coefficient and the road network data of the corresponding road segment, the passability score of each road segment is obtained through the output of a neural network. Determine whether the passability score is greater than a preset threshold; otherwise, remove the road segment and obtain the initial detourable road segments. Among them, the preset threshold is a dynamic variable related to the status data of cold chain goods. The preset threshold refers to the minimum score standard used to screen detour sections, which is realized by dynamic calculation and the basic threshold is adjusted by combining the type of cold chain goods. The dynamic variable refers to the preset threshold changing dynamically according to the preservation requirements of goods, using cold chain goods type data as input parameters. Cold chain goods type data refers to the temperature control requirements classification information of goods during transportation, which is collected by the Internet of Things monitoring system and transmitted to the planning system. Based on the road network data, the predicted values ​​of traffic flow and carrying capacity of each initial detour route are obtained through time-series prediction. Determine whether the predicted value of the carrying capacity data is less than the set value. If so, remove the corresponding part from the initial detourable road section to obtain the detourable road section. Based on the predicted traffic flow and carrying capacity data of the detour routes, the estimated travel time for each detour route is calculated. Detour routes with inconsistent estimated travel times are then eliminated by combining the cold chain cargo status data, resulting in the final detour routes. In calculating the estimated travel time for each detour route, for cold chain cargo, the estimated travel time is not only related to traffic flow and route length, but also closely related to the temperature, humidity, and cargo type status data of the cold chain cargo. By combining cold chain cargo status data, detour routes that do not match the estimated travel time are eliminated to obtain the final detour routes. Specifically, this involves: calculating the minimum remaining shelf life in the cold chain cargo status data; determining whether the estimated travel time is greater than the minimum remaining shelf life, and if so, eliminating the detour route to obtain the final detour route; where the remaining shelf life refers to the maximum time threshold for maintaining the quality and safety of cold chain cargo under the current temperature conditions, which is calculated by collecting cargo temperature data in real time through IoT sensors and combining it with a preset cargo category temperature decay model. Based on the location data, the final detour route, and the estimated travel time, a route is planned and a cold chain transportation route is output.

2. The rapid logistics transportation planning method based on big data analysis according to claim 1, characterized in that: The acquisition of traffic incident data, road network data, location data, and cold chain cargo status data specifically includes: By connecting with a satellite remote sensing data platform, traffic incident data can be obtained in real time. The traffic incident data includes the type of incident, the road section where the incident occurred, the time of occurrence, the scope of impact of the incident, and the duration of the incident. Road network data is obtained from a map service platform. The road network data includes the geographical location, length, real-time traffic data, and maximum carrying capacity of each road segment. Location data is acquired in real time through the transportation management system. The location data includes the vehicle's current geographic coordinates, destination coordinates, and transportation trajectory. The vehicle's Internet of Things (IoT) monitoring system acquires cold chain cargo status data, which includes cargo type, current temperature, and remaining shelf life.

3. The rapid logistics transportation planning method based on big data analysis according to claim 1, characterized in that: The traffic incident data is subjected to nonlinear weighted calculation, and the impact coefficient of the traffic incident data on each road segment is output, specifically including: Multi-dimensional feature extraction is performed on the collected traffic incident data; Multi-dimensional features are processed through a nonlinear weighted model to output the initial impact coefficients of traffic event data on each road segment; the nonlinear weighted model can automatically adjust the weight coefficients of each feature according to the actual impact of different event types on traffic flow and road capacity. The initial impact coefficient of each traffic incident data on each road segment is assigned to the corresponding road segment; For the same road segment affected by multiple traffic incident data, the initial impact coefficient of each traffic incident data is weighted and calculated to obtain the final impact coefficient of the road segment.

4. The rapid logistics transportation planning method based on big data analysis according to claim 1, characterized in that: Based on the aforementioned influence coefficients and the corresponding road network data, the passability score for each road segment is obtained through neural network output, specifically including: The influence coefficients and the corresponding road network data are normalized. A neural network is constructed, which is a graph neural network. Its topology is built based on actual road network data, where nodes represent intersections and edges represent road segments. Traffic impact features of adjacent road segments are aggregated through graph convolutional layers, specifically including: The input layer receives the input of influence coefficients and road network data; At least one hidden layer employs the ReLU activation function for nonlinear transformation and introduces an attention mechanism to dynamically calculate the weights of each input feature; The output layer uses the Sigmoid function to output the passability score of the road segment; Historical traffic event data, historical road network data, and the passability scores of corresponding road segments are used as supervisory signals to train the neural network using supervised learning. The normalized influence coefficients and the corresponding road network data are input into the trained neural network to output the passability score of each road segment.

5. The rapid logistics transportation planning method based on big data analysis according to claim 2, characterized in that: Determining whether the passability score is greater than a preset threshold, and otherwise removing the road segment to obtain the initial detour routes specifically includes: Determine whether the passability score is greater than a preset threshold; otherwise, remove the road segment and obtain the initial detourable road segments. Based on the location data, verify whether the initial detour route satisfies a complete path from the vehicle's current geographical coordinates to the destination coordinates; When the verification result is not met, the initial detour routes are supplemented according to the passability score from high to low until the initial detour routes satisfy the complete path from the vehicle's current geographical coordinates to the destination coordinates. The preset threshold is a dynamic variable that is related to the status data of cold chain goods.

6. The rapid logistics transportation planning method based on big data analysis according to claim 4, characterized in that: Based on the road network data, the predicted traffic flow and carrying capacity data for each initial detour route are obtained through time-series prediction, specifically including: Construct a prediction model that integrates temporal features; Extract historical traffic flow data and historical maximum carrying capacity data of road segments from historical road network data; Historical traffic data and historical maximum carrying capacity data of road segments were converted into supervised learning datasets using the sliding window method. The prediction model is trained using a supervised learning dataset; The road network data is input into the prediction model, and the predicted values ​​of traffic flow and carrying capacity data for each initial detour segment are output.

7. The rapid logistics transportation planning method based on big data analysis according to claim 2, characterized in that: Based on the available detour routes and the corresponding traffic flow predictions, the estimated travel time for each detour route is calculated, specifically including: The length of the detour route is obtained from the road network data corresponding to the detour route. Calculate the traffic flow velocity of the detour route based on the predicted traffic flow data and carrying capacity data of the detour route. The estimated travel time for the detour route is calculated based on the traffic flow speed and length of the detour route.

8. The rapid logistics transportation planning method based on big data analysis according to claim 2, characterized in that: By combining cold chain cargo status data and eliminating alternative routes that do not match the estimated travel time, the final alternative routes include: The minimum remaining shelf life is calculated from the status data of cold chain goods. Determine whether the estimated travel time is greater than the minimum remaining shelf life. If so, remove the detour route and obtain the final detour route.

9. The rapid logistics transportation planning method based on big data analysis according to claim 2, characterized in that: Based on the location data, the final detour route, and the estimated travel time, route planning and output of the cold chain transportation route specifically include: Obtain the vehicle's current geographic coordinates and destination coordinates; The final detour route is used as input data for route planning; The shortest path algorithm is used to calculate the optimal path between the vehicle's current geographical coordinates and the destination coordinates, based on the estimated travel time of each road segment. The optimal path is the path with the lowest estimated travel time, and the estimated travel time is less than the minimum value of the remaining shelf life. Output the optimal path and perform dynamic tracking.

10. A rapid logistics transportation planning system based on big data analysis, characterized in that, include: The data acquisition module is used to acquire traffic event data, road network data, location data, and cold chain cargo status data. Among them, the cold chain cargo status data is acquired through the vehicle's IoT monitoring system. The cold chain cargo status data includes cargo type, current temperature, and remaining shelf life. The Internet of Things (IoT) monitoring system refers to a sensor network deployed in cold chain transport vehicles. It is implemented by combining temperature sensors, timers, and wireless transmission modules to monitor the temperature status of goods and the remaining shelf life. The passability score calculation module is used to perform nonlinear weighted calculation on the traffic incident data and output the impact coefficient of the traffic incident data on each road segment. Based on the influence coefficient and the road network data of the corresponding road segment, the passability score of each road segment is obtained through the output of a neural network. The initial detourable road segment determination module is used to determine whether the passability score is greater than a preset threshold; otherwise, the road segment is removed to obtain the initial detourable road segment. Among them, the preset threshold is a dynamic variable related to the status data of cold chain goods. The preset threshold refers to the minimum score standard used to screen detour sections, which is realized by dynamic calculation and the basic threshold is adjusted by combining the type of cold chain goods. The dynamic variable refers to the preset threshold changing dynamically according to the preservation requirements of goods, using cold chain goods type data as input parameters. Cold chain goods type data refers to the temperature control requirements classification information of goods during transportation, which is collected by the Internet of Things monitoring system and transmitted to the planning system. The detourable road segment identification module is used to obtain the predicted values ​​of traffic flow data and carrying capacity data of each initial detourable road segment through time-series prediction based on the road network data. Determine whether the predicted value of the carrying capacity data is less than the set value. If so, remove the corresponding part from the initial detourable road section to obtain the detourable road section. The final detour route output module calculates the estimated travel time for each detour route based on the available detour routes and the predicted traffic flow and carrying capacity data of the corresponding routes. It also eliminates detour routes with inconsistent estimated travel times by combining the cold chain cargo status data, thus obtaining the final detour routes. In calculating the estimated travel time for each detour route, for cold chain cargo, the estimated travel time is not only related to traffic flow and route length, but also closely related to the temperature, humidity, and cargo type status data of the cold chain cargo. By combining cold chain cargo status data, detour routes that do not match the estimated travel time are eliminated to obtain the final detour routes. Specifically, this involves: calculating the minimum remaining shelf life in the cold chain cargo status data; determining whether the estimated travel time is greater than the minimum remaining shelf life, and if so, eliminating the detour route to obtain the final detour route; where the remaining shelf life refers to the maximum time threshold for maintaining the quality and safety of cold chain cargo under the current temperature conditions, which is calculated by collecting cargo temperature data in real time through IoT sensors and combining it with a preset cargo category temperature decay model. The cold chain transportation route planning module is used to plan routes and output cold chain transportation routes based on the location data, the final detour route, and the estimated travel time.

Images