Large goods transportation path optimization method and system based on road prediction

Abstract

The invention relates to a large goods transportation path optimization method and system based on road prediction, and the method comprises the following steps: obtaining a road information database and road actual condition data, and carrying out the preprocessing of the road actual condition data; based on the road information database and the preprocessed road actual condition data, adopting a pre-trained SVM (Support Vector Machine)-based road prediction model to predict the road condition in a period of time in the future; and based on the road condition and the road information database in the future period of time, performing path planning by adopting a pre-trained path planning model based on deep learning, and outputting a real-time optimal planning path for transportation of the large goods and materials. Compared with the prior art, the method has the advantages of being accurate in road prediction, improving transportation safety and efficiency and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to a method and system for optimizing the transportation path of large materials based on road prediction. Background Art

[0002] In modern society, large electric power materials, such as wind turbine towers and transformers, face many challenges in transportation due to their large size and heavy weight. Traditional transportation methods are often unable to adapt to complex and changeable road conditions, especially in special sections such as narrow roads and sharp bends, which can easily lead to transportation accidents and affect transportation efficiency and safety. Therefore, how to optimize the transportation path of large electric power materials and improve the safety and efficiency of transportation is an important topic of common concern in the current fields of transportation planning and management, logistics and supply chain management, and artificial intelligence and machine learning.

[0003] Solutions of existing technologies: There are two main types of existing solutions: one is the traditional path planning method based on GPS and map data, and the other is the intelligent path planning method based on artificial intelligence and machine learning. The traditional path planning method mainly plans transportation routes through GPS and map data combined with some simple algorithms, such as Dijkstra algorithm, A* algorithm, etc. Although this method can achieve basic route planning, it often cannot provide accurate driving plans when facing special road conditions such as narrow roads and sharp bends.

[0004] Problems with existing technologies: The main problems with existing technologies in this field include:

[0005] First, there is a lack of accurate prediction of the actual road conditions, especially special road conditions such as narrow roads and sharp bends, which cannot be predicted in advance, making it impossible to make decisions based on the actual road conditions;

[0006] Second, although the existing intelligent path planning methods can achieve a certain degree of automation and intelligence, they often cannot provide the best driving plan when faced with complex and changeable road conditions, resulting in low transportation efficiency and even safety hazards.

[0007] Therefore, how to improve the accuracy of road prediction, provide driving plans for narrow and other special roads, and improve the safety and efficiency of transportation are important challenges facing current technology. Summary of the invention

[0008] The purpose of the present invention is to provide a method and system for optimizing the transportation route of large-scale materials based on road prediction to improve the safety of transportation of large-scale materials.

[0009] The purpose of the present invention can be achieved by the following technical solutions:

[0010] A method for optimizing the transportation path of large materials based on road prediction includes the following steps:

[0011] Acquire a road information database and road real-time data, and pre-process the road real-time data;

[0012] Based on the pre-processed road condition data, a pre-trained SVM-based road prediction model is used to predict the road condition in the future.

[0013] Based on the road conditions and road information database in the future period, a pre-trained deep learning-based path planning model is used for path planning to output the real-time optimal planning path for large-scale material transportation.

[0014] Furthermore, the step of constructing the road information database includes:

[0015] Use IoT technology and sensor technology to collect road information, including road information and weather conditions. Each road information includes road width, curve radius, slope, traffic flow, road direction, and road starting and ending points;

[0016] The road information is preprocessed to construct a road information database, and the preprocessing step includes data cleaning, noise removal and outlier removal.

[0017] Furthermore, the training step of the SVM-based road prediction model includes:

[0018] Filtering relevant data for training from the road information database, including road width, curve radius, slope, traffic flow, and weather conditions;

[0019] Construct a road prediction model based on SVM and set the penalty coefficient and kernel parameters of SVM;

[0020] The selected relevant data are input into the SVM-based road prediction model for training, and the penalty coefficient and kernel parameter are adjusted by using the K-fold cross validation method until the optimal parameter combination is reached, thereby obtaining a trained SVM-based road prediction model.

[0021] Furthermore, the road conditions in the future period of time include road traffic congestion, road safety level and road construction conditions in the future period of time.

[0022] Furthermore, the training steps of the deep learning-based path planning model include:

[0023] Collecting driving data of vehicles on the road;

[0024] Acquire multiple road information from the road information database, including image data including road width, curve radius, slope, road direction, road start and end points, and time series data including traffic flow and weather conditions;

[0025] Extract initial features from driving data, image data, and time series data, and perform normalization;

[0026] Constructing a path planning model based on deep learning, selecting a loss function and an optimizer, wherein the path planning model includes two networks, namely a convolutional neural network and a recurrent neural network;

[0027] The normalized initial features of the image data are input into the convolutional neural network, and the normalized initial features of the driving data and the normalized initial features of the time series data are input into the recurrent neural network for training so that the path planning model automatically learns the deep feature representation and obtains a trained path planning model based on deep learning.

[0028] Furthermore, the deep learning-based path planning model is integrated with a reinforcement learning mechanism, and the reinforcement learning mechanism is provided with a state space, an action space and a reward function. The state space includes the current position, road conditions, and vehicle status information; the action space includes selecting the driving direction and adjusting the speed; and the reward function includes driving efficiency and safety.

[0029] Furthermore, it also includes the use of vehicle sensors, positioning systems and high-precision maps to achieve real-time monitoring of the vehicle's operating status, where the real-time monitoring data of the vehicle's operating status includes vehicle speed, acceleration, steering angle, braking status, and GPS location information.

[0030] Furthermore, it also includes transportation risk assessment and early warning steps, including:

[0031] The actual road data and the real-time monitoring data of the vehicle operation status are input into a pre-trained transportation risk assessment and early warning model, the vehicle transportation score is output, the risk in the transportation process is assessed in real time, and an early warning message is issued when a risk is detected.

[0032] Furthermore, the training steps of the transportation risk assessment and early warning model include:

[0033] Obtain road information and historical monitoring data of vehicle operation status, annotate normal and abnormal labels based on historical accident records, and construct a training data set;

[0034] Construct a transportation risk assessment and early warning model based on a neural network combined with an attention mechanism, and select a loss function;

[0035] The training data set is input into the transportation risk assessment and early warning model based on the neural network combined with the attention mechanism, and the parameters are optimized by using the stochastic gradient descent method or its variants until the training conditions are met to obtain a trained transportation risk assessment and early warning model.

[0036] The present invention also provides a large-scale material transportation route optimization system based on road prediction, comprising:

[0037] Road information collection and analysis module: used to collect road information and road real-time data using Internet of Things technology and sensor technology, build a road information database based on the collected road information, and analyze road conditions;

[0038] Road prediction module: used to predict the road conditions in the future period of time based on the road information database and the actual road data by using a pre-trained SVM-based road prediction model;

[0039] Intelligent path planning module: used to perform path planning based on the road conditions and road information database in the future period, using a pre-trained deep learning-based path planning model to output the optimal planned path for the transportation of large-scale materials in real time;

[0040] Vehicle management module: used to use vehicle sensors, positioning systems and high-precision maps to achieve real-time monitoring of vehicle operating status;

[0041] Transport risk assessment and early warning module: used to input the actual road data and real-time monitoring data of vehicle operation status into a pre-trained transport risk assessment and early warning model, output the vehicle transport score, evaluate the risks in the transport process in real time, and issue early warning information based on detected risks.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] (1) The present invention predicts the road conditions in the future through a road prediction model based on SVM, which helps to predict the conditions of untraveled roads in advance. According to the conditions, a deep learning path planning model is used to perform real-time path planning. By combining the future road conditions in this way, the optimal path planning is achieved, thereby improving the safety of the transportation process.

[0044] (2) Due to the large size and heavy weight of large materials, there are many challenges in the transportation process. The present invention takes into account the safety issues of large-item transport vehicles on complex and changeable roads due to special road conditions such as narrow roads and sharp bends through road information such as road width, curve radius, and slope. By predicting various road conditions at future times in advance, not only the accuracy and comprehensiveness of road predictions are improved, but also corresponding preparations can be made in advance to avoid transportation accidents under special road conditions, thereby improving transportation efficiency and safety.

[0045] (3) The path planning model based on deep learning of the present invention is integrated with a reinforcement learning mechanism, which can plan the optimal driving path by continuously learning and optimizing behavior strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0047] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0048] This embodiment provides a method for optimizing the transportation path of large materials based on road prediction. The method is implemented based on a path optimization system, which includes a road information collection and analysis module 1, a road prediction module 2, an intelligent path planning module 3, a vehicle management module 4, and a transportation risk assessment and early warning module 5. According to the system, Figure 1 As shown, the path optimization method includes the following steps:

[0049] Step 1: Road information collection and analysis.

[0050] This step is implemented using a road information collection and analysis module, which uses Internet of Things technology, sensor technology, and big data analysis technology to collect road information, including but not limited to road width, curve radius, slope, traffic flow, etc., and processes and analyzes this data through big data analysis technology to build an accurate road information database.

[0051] Specifically, this step includes the following subdivision steps:

[0052] Step 1.1, using IoT technology, GPS and sensors to collect road information and transmit it to the cloud server. The collected data includes but is not limited to road width, curve radius, slope, traffic flow, etc. The frequency of data collection is once per second, and the scope of the data collection database includes all major highways and urban roads in the country;

[0053] Step 1.2: Clean the collected raw data to remove noise and outliers to ensure the accuracy and reliability of the data;

[0054] Step 1.3: Use data mining, machine learning, and deep learning technologies to conduct in-depth analysis of road information data and extract useful information, including analyzing traffic flow data to predict traffic congestion, analyzing road condition data to assess road safety levels, and analyzing road construction conditions for subsequent route planning;

[0055] Step 1.4, Visualization: Visualize the analysis results in the form of charts, maps, etc. to facilitate traffic managers and the public to intuitively understand road information.

[0056] Step 2: Road prediction.

[0057] This step is implemented using a road prediction module. The road prediction module uses machine learning technology to establish a road prediction model based on road information data to predict road conditions in the future, especially for special road conditions such as narrow roads and sharp bends, so that measures can be taken in advance to ensure transportation safety. The road prediction model is established using a support vector machine (SVM). During the model training process, the cross-validation method is used for parameter optimization, the penalty coefficient is set to 1, the kernel parameter is set to 2, and the number of training rounds is 1000.

[0058] Specifically, road prediction includes the following steps:

[0059] Step 2.1, filter out relevant data for prediction model training from the road information database, especially data for special road conditions such as narrow roads and sharp bends. Clean the filtered data, including processing missing values ​​and outliers, and necessary normalization or standardization to ensure data quality. According to the prediction target (such as road congestion level, safety level, etc.), extract and construct feature sets such as historical traffic flow, road width, bend radius, slope, weather conditions, etc.

[0060] Step 2.2, based on the characteristics of the problem and the advantages of support vector machine (SVM) in handling classification and regression problems, select SVM as the basis of the prediction model. Preliminary set the penalty coefficient (C = 1) and kernel parameters (such as γ = 2 of RBF kernel) of SVM. Use K-fold cross validation (such as K = 5 or 10) to evaluate the model performance and optimize the parameters accordingly.

[0061] Step 2.3, divide the preprocessed data into training set and test set, usually according to a certain ratio (such as 70% training set, 30% test set). Use the training set data to train the SVM model, set the number of training rounds to 1000 times, and ensure that the model fully learns the data features. During the training process, use the results of cross-validation to adjust the penalty coefficient and kernel parameters of the SVM to find the optimal combination of model parameters.

[0062] Step 2.4, use the test set data to evaluate the trained model and calculate performance indicators such as accuracy, recall, F1 score, etc. Analyze the difference between the model prediction results and the actual values, and identify possible sources of error, such as data noise, improper feature selection, etc. According to the evaluation results and error analysis, adjust the feature set, improve the data preprocessing method, etc. to make necessary adjustments and optimizations to the model.

[0063] In practical applications, by collecting actual road data and inputting it into the trained road prediction model, the road conditions in the future can be predicted, including road traffic congestion, road safety level and road construction conditions in the future.

[0064] Step 3: Intelligent path planning.

[0065] This step is implemented using the intelligent path planning module. The intelligent path planning module develops a path planning model based on deep learning and reinforcement learning. By analyzing a large amount of historical transportation data, it learns the optimal driving strategy and dynamically adjusts the path planning plan based on real-time road information. It provides the optimal driving plan and improves transportation efficiency and safety.

[0066] Specifically, this step includes the following sub-steps:

[0067] Step 3.1, in order to improve the efficiency of large-scale material transportation and reduce driving time, collect vehicle driving data, including vehicle driving speed, driving trajectory, etc., obtain multiple road information from the road information database, including image data such as road width, curve radius, slope, road direction, road start and end points, and time series data such as traffic flow and weather conditions. These data constitute a multi-dimensional historical data set;

[0068] Step 3.2: Remove noise data, outliers, and duplicate records from the collected data to ensure data quality, and extract useful features from the original data, such as the coordinates of the start and end points, timestamps, average speed of the road section, weather codes, etc. Standardize or normalize the features to eliminate the impact of different dimensions on model training. Divide the preprocessed data into training set, validation set, and test set.

[0069] Step 3.3, introduce convolutional neural network and recurrent neural network, design a deep learning model structure suitable for path planning problems, integrate reinforcement learning mechanism on the basis of deep learning model, and use TensorFlow framework to implement the above designed algorithm. The details are as follows:

[0070] Step 3.3.1, introduce convolutional neural network and recurrent neural network, design a deep learning model structure suitable for the path planning problem, and input the data classification preprocessed in step 3.2 into the deep learning model. Specifically, input the image data into the convolutional neural network and the time series data into the recurrent neural network, so that the deep learning model automatically learns the deep feature representation of the input data.

[0071] Step 3.3.2, based on the deep learning model, the reinforcement learning mechanism is integrated to define the state space, action space and reward function. The state space can include information such as the current location, road conditions, and vehicle status; the action space can include decisions such as choosing the driving direction and adjusting the speed; and the reward function is designed based on factors such as driving efficiency and safety.

[0072] Step 3.3.3, use the TensorFlow framework to implement the above designed algorithm, including model definition, loss function design, optimizer selection (such as stochastic gradient descent method, learning rate set to 0.001), etc.

[0073] Step 3.4, set the initial parameters for model training, such as learning rate, number of training rounds (5000 times), batch size, etc. Use the training set data to train the model and continuously optimize the model parameters by stochastic gradient descent. Evaluate the model performance on the validation set and adjust the model structure, parameters or optimization algorithm based on the validation results. Set early stopping conditions to avoid overfitting and save the best performing model during training. Deploy the trained model to the server or vehicle-mounted device to ensure real-time response.

[0074] Step 3.5 ensures that the algorithm can access real-time road information (such as traffic flow, accident reports, weather warnings, etc.) in order to dynamically adjust the path planning solution.

[0075] After training, a trained path planning model is obtained. By inputting the road conditions in the future and the road information in the road information database into it, the path planning is carried out in real time and optimally.

[0076] Step 4: Vehicle management.

[0077] This step is implemented using the vehicle management module, which performs intelligent management of large-scale electric power material transportation vehicles, including vehicle sensors, positioning systems and high-precision maps, to achieve real-time monitoring of vehicle location, speed, posture and other information, and automatically adjust driving strategies based on the guidance of intelligent path planning algorithms to improve the degree of automation and safety of transportation.

[0078] Step 5: Transportation risk assessment and early warning.

[0079] This step is implemented using the transportation risk assessment and early warning module. The transportation risk assessment and early warning module: based on real-time road data and vehicle operation status monitoring data, establishes a transportation risk assessment and early warning system to assess the risks that may be faced during transportation in real time, and when potential risks are discovered, promptly issues early warning information to remind drivers to take appropriate measures to avoid accidents.

[0080] Specifically, this step includes the following sub-steps:

[0081] Step 5.1, choose a popular deep learning framework such as TensorFlow or PyTorch to develop a neural network model.

[0082] Step 5.2: For the road prediction model data such as historical traffic flow, road conditions (such as construction, accidents), weather conditions, and vehicle operation status monitoring data such as vehicle speed, acceleration, steering angle, brake status, GPS location information, etc., normal and abnormal behavior data are annotated according to historical accident records.

[0083] Step 5.3, design a neural network structure suitable for anomaly detection, such as autoencoder, long short-term memory network (LSTM) combined with attention mechanism, etc., to capture the complex characteristics of time series data. Choose a loss function suitable for anomaly detection, such as reconstruction error (for autoencoder) or probability-based loss function (such as negative log-likelihood loss).

[0084] In step 5.4, use stochastic gradient descent SGD or its variants (such as Adam, RMSprop) to optimize the model parameters. For example, specify a learning rate of 0.0001, which is a relatively small value that helps the model to learn stably. Set the number of training rounds to 10,000 to ensure that the model has enough time to learn from the data and converge to better performance. To prevent overfitting, an early stopping mechanism can be introduced to stop training early when the performance on the validation set does not improve significantly over multiple consecutive rounds.

[0085] Step 5.5: Ensure that the system can receive and process data from vehicle and road monitoring systems in real time. Based on the anomaly scores output by the model, the risks in the transportation process are assessed in real time, and when potential risks are detected, early warning information is sent to drivers or managers through vehicle terminals, mobile phone apps, or text messages.

[0086] Based on the above modules, a transportation dispatching and management system based on cloud computing technology is constructed. The system includes a road information collection and analysis module, a road prediction module, an intelligent path planning module, a vehicle management module, and a transportation risk assessment and early warning module. It can realize the whole process monitoring and management of the transportation of large-scale electric power materials, optimize the allocation of large-scale material transportation resources, and improve the efficiency and benefits of transportation.

[0087] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0088] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0089] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0092] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0093] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for optimizing the transportation path of large materials based on road prediction, characterized in that: The following steps are involved: Acquire a road information database and road real-time data, and pre-process the road real-time data; Based on the pre-processed road condition data, a pre-trained SVM-based road prediction model is used to predict the road condition in the future. Based on the road conditions and road information database in the future period, a pre-trained deep learning-based path planning model is used for path planning to output the real-time optimal planning path for large-scale material transportation.

2. The method for optimizing the transportation path of large materials based on road prediction according to claim 1 is characterized in that: The steps of constructing the road information database include: Use IoT technology and sensor technology to collect road information, including road information and weather conditions. Each road information includes road width, curve radius, slope, traffic flow, road direction, and road starting and ending points; The road information is preprocessed to construct a road information database, and the preprocessing step includes data cleaning, noise removal and outlier removal.

3. The method for optimizing the transportation path of large materials based on road prediction according to claim 1 is characterized in that: The training steps of the SVM-based road prediction model include: Filtering relevant data for training from the road information database, including road width, curve radius, slope, traffic flow, and weather conditions; Construct a road prediction model based on SVM and set the penalty coefficient and kernel parameters of SVM; The selected relevant data are input into the SVM-based road prediction model for training, and the penalty coefficient and kernel parameter are adjusted by using the K-fold cross validation method until the optimal parameter combination is reached, thereby obtaining a trained SVM-based road prediction model.

4. The method for optimizing the transportation path of large materials based on road prediction according to claim 1 is characterized in that: The road conditions in the future period include road traffic congestion, road safety level and road construction conditions in the future period.

5. The method for optimizing the transportation path of large materials based on road prediction according to claim 1 is characterized in that: The training steps of the deep learning-based path planning model include: Collecting driving data of vehicles on the road; Acquire multiple road information from the road information database, including image data including road width, curve radius, slope, road direction, road start and end points, and time series data including traffic flow and weather conditions; Extract initial features from driving data, image data, and time series data, and perform normalization; Constructing a path planning model based on deep learning, selecting a loss function and an optimizer, wherein the path planning model includes two networks, namely a convolutional neural network and a recurrent neural network; The normalized initial features of the image data are input into the convolutional neural network, and the normalized initial features of the driving data and the normalized initial features of the time series data are input into the recurrent neural network for training so that the path planning model automatically learns the deep feature representation and obtains a trained path planning model based on deep learning.

6. The method for optimizing the transportation path of large materials based on road prediction according to claim 5 is characterized in that: The deep learning-based path planning model is integrated with a reinforcement learning mechanism, which is provided with a state space, an action space and a reward function. The state space includes the current position, road conditions, and vehicle status information; the action space includes selecting the driving direction and adjusting the speed; and the reward function includes driving efficiency and safety.

7. The method for optimizing the transportation path of large materials based on road prediction according to claim 1 is characterized in that: It also includes the use of vehicle sensors, positioning systems and high-precision maps to achieve real-time monitoring of the vehicle's operating status, where the real-time monitoring data of the vehicle's operating status includes vehicle speed, acceleration, steering angle, braking status, and GPS location information.

8. The method for optimizing the transportation path of large materials based on road prediction according to claim 1 is characterized in that: It also includes transportation risk assessment and early warning steps, including: The actual road data and the real-time monitoring data of the vehicle operation status are input into a pre-trained transportation risk assessment and early warning model, the vehicle transportation score is output, the risk in the transportation process is assessed in real time, and an early warning message is issued when a risk is detected.

9. The method for optimizing the transportation path of large materials based on road prediction according to claim 8 is characterized in that: The training steps of the transportation risk assessment and early warning model include: Obtain road information and historical monitoring data of vehicle operation status, annotate normal and abnormal labels based on historical accident records, and construct a training data set; Construct a transportation risk assessment and early warning model based on a neural network combined with an attention mechanism, and select a loss function; The training data set is input into the transportation risk assessment and early warning model based on the neural network combined with the attention mechanism, and the parameters are optimized by using the stochastic gradient descent method or its variants until the training conditions are met to obtain a trained transportation risk assessment and early warning model.

10. A large material transportation route optimization system based on road prediction, characterized in that: include: Road information collection and analysis module: used to collect road information and road real-time data using Internet of Things technology and sensor technology, build a road information database based on the collected road information, and analyze road conditions; Road prediction module: used to predict the road conditions in the future period of time based on the road information database and the actual road data by using a pre-trained SVM-based road prediction model; Intelligent path planning module: used to perform path planning based on the road conditions and road information database in the future period, using a pre-trained deep learning-based path planning model to output the optimal planned path for the transportation of large-scale materials in real time; Vehicle management module: used to use vehicle sensors, positioning systems and high-precision maps to achieve real-time monitoring of vehicle operating status; Transport risk assessment and early warning module: used to input the actual road data and real-time monitoring data of vehicle operation status into a pre-trained transport risk assessment and early warning model, output the vehicle transport score, evaluate the risks in the transport process in real time, and issue early warning information based on detected risks.

Images