Intelligent sensing and active early warning system and method for safety risk of multimodal transport of long and large goods based on multi-source data fusion

Abstract

The invention, which relates to the technical field of logistics transportation and intelligent transportation, discloses an intelligent sensing and active early warning system for security risks of multimodal transport of long and large goods based on multi-source data fusion, comprising: an original data input module; a service scene intelligent sensing module; an acquired data analysis module; a safety risk early warning module; an operation guidance module; a networking information receiving module; the invention further discloses a long and large cargo multimodal transport safety risk intelligent sensing and active early warning method based on multi-source data fusion. According to the invention, the complex safety challenge faced by long and large goods in the multimodal transportation process is solved, and the safety and efficiency of the transportation process are improved.

Description

Technical Field

[0001] The present invention relates to the field of logistics, transportation and intelligent traffic technology, and in particular to an intelligent perception and active early warning system and method for safety risks of long and large multimodal transport of goods based on multi-source data fusion. Background Art

[0002] Long and bulky cargo is a general term for oversized road freight, overweight and oversize rail freight, and bulky and oversized water freight. These cargoes are characterized by being bulky, large, and expensive. The transportation of these long and bulky cargoes is a crucial component of major national projects, such as those involving electric power, petrochemicals, and other key projects. It directly impacts the construction progress and safety of major infrastructure projects, including thermal power, hydropower, nuclear power, petrochemicals, and smelting. These long and bulky cargoes primarily include large power equipment (transformers, generator stators, wind turbine blades and towers, etc.), large manufacturing equipment (machining equipment, production line components, molds, etc.), and large construction machinery (excavators, loaders, bulldozers, rollers, cranes, etc.). These cargoes are typically large in size and heavy, placing extremely stringent requirements on transportation conditions. With the rapid development of industrialization and modernization, the demand for transporting these long and bulky cargoes is also growing.

[0003] However, transporting long and bulky cargo presents numerous challenges. First, due to limitations on equipment size and weight, special attention must be paid to the width, height, and load capacity of roads, as well as the passability of bridges and tunnels. Complex transport scenarios, such as curves on city streets, low spaces within tunnels, narrow passages in culverts, and the load limitations of bridges, can all pose obstacles during transportation. These scenarios can easily lead to safety incidents such as collisions, over-limit, and overloading. Once these occur, they can not only cause significant property damage but also pose a threat to personal safety. Furthermore, the transport of long and bulky cargo is also affected by meteorological conditions. For example, severe weather conditions such as strong winds, heavy rain, snow, and fog can impact transportation safety and increase transportation risks. Therefore, real-time monitoring of weather changes is necessary during transportation, and transportation plans and strategies must be adjusted accordingly.

[0004] Existing transportation monitoring systems often fail to meet these complex requirements. They only provide limited data, such as vehicle location and speed, and lack comprehensive awareness of the transportation environment, making it impossible to monitor entire road sections. Furthermore, these systems often lack sufficient data analysis capabilities, making it difficult to identify potential safety risks from large amounts of data. The delay in early warning systems is also a problem, with alerts often issued only after an incident has occurred, by which time losses can no longer be avoided.

[0005] Therefore, developing a new intelligent safety risk perception and early warning system is crucial for the transportation of large and heavy cargo. This system needs to be able to perceive the transportation environment in real time, including the operating status of the transportation vehicle, the securing status of the large and heavy cargo, the operating behavior of relevant personnel, road conditions, climate change, and other factors. It must also be able to intelligently analyze this data to identify potential risk factors. More importantly, the system needs to provide early warning, giving transportation personnel ample time to take action and avoid accidents. This will not only improve transportation safety, but also increase transportation efficiency and reduce transportation costs, providing strong support for the development of major national projects. Summary of the Invention

[0006] To address the problems existing in the prior art, the present invention aims to provide an intelligent perception and active warning system and method for safety risks in the multimodal transport of long and large cargoes based on multi-source data fusion. The present invention addresses the complex safety challenges faced by long and large cargoes in the multimodal transport process and improves the safety and efficiency of the transportation process.

[0007] To achieve the above objectives, the present invention adopts a technical solution: an intelligent perception and active early warning system for safety risks of long and large multimodal transport of goods based on multi-source data fusion, comprising:

[0008] Raw data input module: used to receive user-entered detailed information on the dimensions, weight, transportation route, personnel information, and transportation vehicle of large and large cargo, and to establish an initial data model to provide a basis for intelligent analysis; and also used to import historical transportation data and external data sources;

[0009] Business scenario intelligent perception module: Integrates high-definition cameras, Beidou positioning, inclination sensors, acceleration monitors, tension monitors, tire pressure monitors, vehicle level monitors, and a clearance laser rangefinder. It is used to perceive the operating status of transportation vehicles, cargo securing status, personnel operation behavior, and dynamic road environment information in real time; and is used to connect with other intelligent devices to obtain transportation environment information.

[0010] Data Collection and Analysis Module: This module uses big data processing technology and artificial intelligence algorithms to conduct in-depth analysis of data collected by the business scenario intelligent perception module. This includes the compatibility of the vehicle's location, speed, and direction with the preset route; the vehicle's current condition and its adaptability to the road environment; the stability of the cargo's reinforcement status; and the compliance of personnel operations. The module also continuously optimizes analysis models and early warning strategies based on historical data and real-time feedback.

[0011] Safety risk warning module: used to alert the on-site commander and driver when potential safety hazards are discovered; and used for remote monitoring and emergency communication to contact relevant departments and personnel in an emergency;

[0012] Operational Guidance Module: This module combines early warning information to provide operational guidance to on-site transport commanders and drivers. The guidance includes specific emergency measures and operating procedures, as well as explanations of potential risks and preventive measures. It also features a voice prompt function to provide timely guidance when the driver's vision is obstructed or their attention is distracted.

[0013] Network information receiving module: It is connected to the external system to receive and process the latest traffic information and weather information in real time. Based on the received information, it can automatically adjust the early warning strategy and operation guidance plan to deal with emergencies and complex environments.

[0014] The present invention also provides a method for intelligently sensing and proactively warning safety risks of long and large cargo intermodal transport based on multi-source data fusion, which is implemented using the above-mentioned intelligently sensing and proactively warning safety risks of long and large cargo intermodal transport based on multi-source data fusion system. The method comprises the following steps:

[0015] Step 1. System initialization and data input: After logging into the system, the user first enters the detailed dimensions, precise weight, and specific transportation route information of the long and large cargo through the user interface. Subsequently, the input information is automatically verified using the built-in verification algorithm to ensure the integrity and rationality of the data;

[0016] Step 2: Real-time data perception and collection: During transportation, the business scenario intelligent perception module continuously perceives the vehicle's operating status in real time;

[0017] Step 3: Data analysis and risk assessment: After the data analysis module receives the real-time data collected by the perception module, it uses big data processing technology and artificial intelligence algorithms to conduct in-depth analysis;

[0018] Step 4: Early warning and operation guidance: Through multi-channel early warning, ensure that the driver and operator can receive early warning information in time even if they are not in the cab;

[0019] Step 5: Network information reception and processing: Receive official information from traffic management departments and obtain relevant data through interfaces with other traffic information systems;

[0020] Step 6: System maintenance and upgrade: Take various measures to maintain continuous optimization of its performance and functions.

[0021] As a further improvement of the present invention, the step 1 is specifically as follows:

[0022] The system constructs an initial data model based on the verified data, including the description of the equipment shape and size model through a three-dimensional spatial geometry algorithm, and the generation of an optimal or feasible path transportation route model based on the GIS geographic information system and path planning algorithm. At the same time, the system automatically connects and initializes various sensor devices through the built-in communication protocol, including vehicle operation status sensors, environmental perception sensors, cargo reinforcement status monitoring sensors, and personnel operation behavior monitoring sensors.

[0023] As a further improvement of the present invention, in step 2, for monitoring key vehicle parameters, a sensor data acquisition algorithm is used to ensure data accuracy and real-time performance; Beidou positioning information is obtained based on a satellite positioning algorithm to calculate vehicle position, speed, and direction information in real time;

[0024] High-definition cameras and radars use image processing and radar detection algorithms to capture real-time information about the surrounding environment, providing the system with data on road width, height limits, and turning angles. By comparing the vehicle's operating status with road environment information, an adaptability assessment algorithm is used to assess the vehicle's compatibility with the environment in real time, ensuring a safe and smooth transportation process.

[0025] When monitoring the cargo reinforcement status, the lashing tension monitoring and impact vibration detector uses mechanical models and signal processing algorithms to monitor the stress and vibration status of the cargo in real time, and immediately triggers the early warning mechanism when an abnormality is found; at the same time, facial recognition monitoring, fatigue recognition, and distraction recognition equipment continuously monitor personnel's operating behavior through deep learning algorithms to ensure that drivers and operators are in good condition.

[0026] As a further improvement of the present invention, in step 3, big data processing technology includes data cleaning, data aggregation, and data visualization; artificial intelligence algorithms include machine learning algorithms and deep learning algorithms; in terms of the matching analysis of vehicle position, speed, direction and preset route, a path matching algorithm is used to evaluate the degree of deviation between the vehicle's current position and the preset route, and adjustment suggestions are given in combination with road condition information; the adaptability assessment of the vehicle condition and the road environment is carried out through a comprehensive evaluation algorithm, using fuzzy logic or Bayesian network methods to calculate the adaptability score.

[0027] When evaluating the stability of cargo reinforcement status, a cargo status model is constructed through real-time monitoring data, and a status prediction algorithm is used to predict the future status of the cargo. When an abnormality is found, an early warning mechanism is triggered. The compliance assessment of personnel operation behavior is based on the behavior recognition results of the deep learning algorithm, combined with the operation specification library to make compliance judgments. When an illegal operation is found, an alarm is immediately issued and operational guidance suggestions are provided.

[0028] Using a collision risk assessment algorithm, we comprehensively consider multiple factors including vehicle status, road environment, cargo reinforcement status, and personnel operating behavior to accurately determine whether there are potential safety threats in the current transportation process, and provide important basis for subsequent early warning and operational guidance.

[0029] As a further improvement of the present invention, in step 4, the multi-channel warning includes not only sound and light alarms, but also SMS, email or mobile application push notifications; and:

[0030] Interactive display: The in-cab display not only displays warning messages but also provides detailed map views and animated simulations, visually illustrating potential safety hazards and their potential impact, helping personnel better understand and take appropriate measures.

[0031] Voice Command: The system supports voice command function, which automatically plays preset voice warnings or operating instructions for use when the driver's vision is limited or his hands are busy operating.

[0032] The operational guidance module provides personalized risk avoidance advice or operational guidance based on the specific content of the warning. This module is based on the prediction results of big data analysis and machine learning models, and takes into account the current traffic conditions, vehicle status, cargo status, and personnel status.

[0033] Dynamic Path Planning: When the vehicle is about to deviate from the preset route, it advises the driver to slow down or adjust the driving direction. It also dynamically plans a new optimal path based on real-time traffic information and displays it in real time on the cab display screen.

[0034] Cargo reinforcement suggestions: For abnormal cargo reinforcement status, specific reinforcement measures are provided based on the nature, size, weight and stress conditions of the cargo during transportation;

[0035] Personnel operation guidance: Alerts are issued for illegal operations or unsafe behaviors of drivers and operators, and operation guidance videos or graphic descriptions are provided to help personnel correct incorrect behaviors and improve operational standardization;

[0036] After each warning and operation guidance, user feedback information and actual processing results are collected, and the warning model and operation guidance strategy are optimized through machine learning algorithms to improve the accuracy and effectiveness of the system.

[0037] The intelligent perception and early warning system based on multi-source data fusion of long and large cargo multimodal transport of the present invention has brought significant technical and economic effects, and has had a profound positive impact on the transportation of long and large cargo in the power industry.

[0038] By integrating advanced sensor technology, big data processing capabilities, and artificial intelligence algorithms, this system significantly improves the safety and efficiency of long-haul cargo transportation. The system's real-time monitoring capabilities allow it to monitor even the smallest changes during transportation, providing timely warnings of potential risks such as collisions, over-gauge and overloads. This real-time sensing and intelligent analysis capability not only reduces the probability of accidents but also improves adaptability to complex transportation environments, ensuring smooth transportation.

[0039] The implementation of this invention helps reduce transportation costs and improve transportation efficiency. By reducing the occurrence of accidents, it directly reduces repair costs, cargo damage costs, and possible downtime costs caused by accidents. Furthermore, the system's early warning function helps optimize transportation routes and times, reducing unnecessary delays, thereby improving transportation efficiency. Furthermore, the intelligent system reduces reliance on manual monitoring, further reducing labor costs.

[0040] In summary, this invention not only improves the safety and efficiency of long-haul freight transportation at a technical level, but also brings significant cost savings and efficiency improvements to transportation companies at an economic level. These benefits not only enhance the competitiveness of transportation companies but also provide solid logistical support for infrastructure development across the entire power industry.

[0041] The beneficial effects of the present invention are:

[0042] 1. Significantly improve transportation safety: Through real-time perception, data analysis and early warning mechanisms, the present invention can promptly detect and warn of potential safety hazards, effectively reducing the accident rate during the transportation of long and large cargoes.

[0043] 2. Improve transportation efficiency: Intelligent monitoring and early warning systems can optimize transportation routes and speed control strategies, reducing transportation delays caused by avoiding obstacles or waiting for road conditions to improve.

[0044] 3. Reduce operating costs: By reducing the occurrence of accidents and delays, the present invention helps to reduce the operating costs of transportation companies and improve economic benefits.

[0045] 4. Enhanced user experience: Detailed operational guidance and risk avoidance suggestions are provided to enable drivers to more easily cope with complex transportation environments, improving user experience and satisfaction.

[0046] 5. Promote the intelligent development of the logistics industry: The successful application of this invention will set an example for the intelligent transformation of the logistics industry and promote the development of the entire industry towards a more efficient, safe and intelligent direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a system framework diagram of an embodiment of the present invention;

[0048] Figure 2 A flow chart of a method according to an embodiment of the present invention

[0049] Figure 3 This is a schematic representation of the data analysis and risk assessment results in the embodiments of the present invention;

[0050] Figure 4 Flowchart of data analysis and risk assessment in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0052] Example 1

[0053] like Figure 1 As shown in the figure, an intelligent perception and active warning system for safety risks of long and large cargo multimodal transport based on multi-source data fusion includes:

[0054] The Raw Data Input Module receives key user-entered information, including detailed dimensions, weight, transportation routes, personnel information, and transportation vehicle information for large and large shipments. It then builds an initial data model, providing the foundation for subsequent intelligent analysis. This module also supports the import of historical transportation data and external data sources to enhance the accuracy and comprehensiveness of analysis.

[0055] The intelligent perception module for business scenarios integrates multiple sensors, including high-definition cameras, Beidou positioning, inclination sensors, accelerometers, tension monitors, tire pressure monitors, vehicle level monitors, and a laser rangefinder. This module provides real-time perception of dynamic information, including the operating status of transport vehicles, cargo securing status, operator behavior, and road conditions. Furthermore, the module supports seamless integration with other intelligent devices, such as drone inspections and ground mobile monitoring stations, to provide more comprehensive information about the transportation environment.

[0056] The Data Collection and Analysis Module utilizes big data processing technology and artificial intelligence algorithms to conduct in-depth analysis of data collected by the perception module. This analysis covers multiple dimensions, including the compatibility of the vehicle's location, speed, and direction with the pre-set route; the vehicle's current condition with the road environment; the stability of the cargo's reinforcement; and the compliance of personnel operations. This module also possesses self-learning capabilities, continuously optimizing analysis models and early warning strategies based on historical data and real-time feedback.

[0057] Safety Risk Warning Module: Upon detecting potential safety hazards, such as potential collision risk, non-compliance with forward clearance requirements, improper human operation, or cargo status data deviation exceeding a threshold, the Safety Risk Warning Module will immediately trigger an alert to the on-site commander and driver through various means, including sound, lighting, and display screens. The module also supports remote monitoring and emergency communication, ensuring timely contact with relevant departments and personnel in emergencies.

[0058] Operational Guidance Module: This module integrates warning information to provide operational guidance to on-site transport commanders and drivers. This guidance includes not only specific emergency measures and operational procedures, but also explanations of potential risks and preventative measures. This module also features voice prompts to provide timely guidance when the driver's vision is obstructed or their attention is distracted.

[0059] Networked Information Receiving Module: This module connects to external systems such as traffic management and meteorological authorities to receive and process the latest traffic information (such as changes in bridge load limits and temporary road closures) and weather information (such as warnings of severe weather conditions like rain, snow, and fog) in real time. Based on this information, the system automatically adjusts its warning strategies and operational guidance to respond to emergencies and complex environments.

[0060] like Figure 2 As shown, this embodiment also provides a method for intelligent perception and active early warning of safety risks of long and large cargo multimodal transport based on multi-source data fusion, comprising the following steps:

[0061] Step 1: System initialization and data input:

[0062] During this phase, after logging into the system, users first enter the detailed dimensions (e.g., length, width, height) and precise weight of their oversized cargo through the user interface (UI), along with specific transportation route information (including starting point, destination, transit points, and specific road conditions). The system receives this data in structured data formats such as JSON or XML to ensure accuracy and parsability. The system then automatically verifies the input using built-in validation algorithms (e.g., regular expression matching and range checking) to ensure data integrity and validity.

[0063] The system builds an initial data model based on the verified data. These models include but are not limited to equipment size models (describing the shape and size of the equipment through three-dimensional spatial geometry algorithms) and transportation route models (generating optimal or feasible paths based on GIS geographic information systems and path planning algorithms). At the same time, the system automatically connects and initializes various sensor devices through built-in communication protocols (such as CAN bus, MQTT, etc.), such as vehicle operation status sensors (monitoring acceleration, tire pressure, vehicle plate level, etc., and extracting key parameters using sensor data analysis algorithms), environmental perception sensors (combining high-definition cameras with radars, using image processing and radar detection algorithms to measure road width, height limits, etc.), cargo reinforcement status monitoring sensors (binding tension monitoring, impact and vibration detection, and evaluating cargo status through mechanical models and signal processing algorithms), and personnel operation behavior monitoring sensors (face recognition, fatigue recognition, distraction recognition, and real-time monitoring and analysis based on deep learning algorithms).

[0064] Step 2: Real-time data perception and collection:

[0065] During transportation, the intelligent sensing module continuously monitors the vehicle's operating status in real time. For monitoring key vehicle parameters such as acceleration and tire pressure, the system uses sensor data acquisition algorithms (such as ADC analog-to-digital conversion and filtering algorithms to remove noise) to ensure data accuracy and real-time performance. Beidou positioning information is acquired through satellite positioning algorithms, which calculate vehicle position, speed, direction, and other information in real time.

[0066] Advanced equipment such as high-definition cameras and radars capture real-time information about the surrounding environment through image processing algorithms (such as edge detection, feature extraction, and target recognition) and radar detection algorithms (such as distance measurement, speed estimation, and target tracking), providing the system with key data such as road width, height restrictions, and turning angles. By comparing vehicle operating status (such as vehicle size and driving trajectory) with road environment information and utilizing adaptability assessment algorithms (such as collision warning models and path feasibility analysis), the system assesses the compatibility of the vehicle with the environment in real time, ensuring safe and smooth transportation.

[0067] In terms of cargo reinforcement status monitoring, lashing force monitoring and impact vibration detectors use mechanical models and signal processing algorithms (such as Fast Fourier Transform (FFT) to analyze vibration frequency and amplitude, and stress-strain analysis to assess tension) to monitor the cargo's stress and vibration conditions in real time, triggering warning mechanisms immediately upon detecting anomalies. Meanwhile, facial recognition monitoring, fatigue detection, and distraction detection equipment use deep learning algorithms (such as convolutional neural networks (CNN) for facial feature extraction and recurrent neural networks (RNN) for behavioral pattern analysis) to continuously monitor operator behavior and ensure the health of drivers and operators.

[0068] Step 3: Data Analysis and Risk Assessment:

[0069] After receiving the real-time data collected by the perception module, the data analysis module uses big data processing technology and artificial intelligence algorithms to conduct in-depth analysis. Big data processing technology includes steps such as data cleaning (removing outliers and filling missing values), data aggregation (aggregating data by time, space, category, and other dimensions), and data visualization (generating charts, heat maps, and other intuitive displays of data). Artificial intelligence algorithms include machine learning algorithms (such as classification, regression, and clustering) and deep learning algorithms (such as convolutional neural networks (CNN) and long short-term memory networks (LSTM). These algorithms can automatically identify key features and patterns in data, providing the system with comprehensive data insights.

[0070] To analyze the compatibility of a vehicle's position, speed, and direction with the preset route, the system uses path-matching algorithms (such as Hausdorff distance and Frechet distance) to assess the degree of deviation between the vehicle's current position and the preset route, and then provides adjustment recommendations based on road condition information. The system also uses a comprehensive evaluation algorithm (taking into account vehicle performance parameters, road conditions, weather factors, etc.) to calculate an adaptability score using methods such as fuzzy logic or Bayesian networks.

[0071] Regarding cargo reinforcement stability assessment, the system constructs cargo status models (such as finite element analysis models and dynamic mechanics models) from real-time monitoring data. It then uses state prediction algorithms (such as Kalman filtering and particle filtering) to predict the cargo's future state and trigger early warning mechanisms when anomalies are detected. The compliance assessment of personnel operations is based on the behavioral recognition results of deep learning algorithms, combined with a database of operational specifications to determine compliance. When violations are detected, an alert is immediately issued and operational guidance is provided.

[0072] Based on the above analysis results, the system uses a collision risk assessment algorithm (Monte Carlo simulation) to comprehensively consider multiple factors such as vehicle status, road environment, cargo reinforcement status, and personnel operating behavior to accurately determine whether there are potential safety threats in the current transportation process, and provide important basis for subsequent early warning and operational guidance.

[0073] Specific steps:

[0074] (1) Data cleaning: First, the collected data is cleaned to remove outliers and fill missing values. For example, the mean is used to fill missing values ​​in numerical data, and outliers are identified and removed or replaced with the median. Suppose there are 1000 data points, of which 5 speed data points are missing and 3 are outliers. The missing values ​​are filled by calculating the mean, and the outliers in speed are replaced with the median.

[0075] (2) Data aggregation: Aggregate the cleaned data. For example, based on the time dimension, aggregate data within every 10 minutes into a group to analyze the changing trends of vehicle driving status. In the spatial dimension, data can be classified and summarized according to different road sections. In the category dimension, data can be divided into categories such as vehicle status, road environment, and cargo status for analysis.

[0076] (3) Data visualization: Generate charts and heat maps to display data intuitively. For example, draw a line graph showing vehicle position over time, generate a histogram of vehicle speed distribution, and generate a heat map of the frequency of road accidents.

[0077] (4) Path matching algorithm: Use Hausdorff distance or Frechet distance to calculate the degree of deviation between the vehicle's current position and the preset route. For example, set the Hausdorff distance threshold to 0.1 km. If the calculated deviation distance is greater than the threshold, it is considered a significant deviation.

[0078] Formula example:

[0079]

[0080] (5) Comprehensive evaluation algorithm: Considering vehicle performance parameters, road conditions, weather factors, etc., fuzzy logic or Bayesian network is used to calculate the adaptability score. For example, fuzzy logic is used to evaluate vehicle performance, set weight factors, and calculate the comprehensive score.

[0081] Formula example:

[0082] S=w1·P1+w2·P2+…+w n ·P n

[0083] Where S is the adaptability score, w i is the weight factor, P i It is a performance parameter.

[0084] Specifically, considering vehicle performance parameters, road conditions, weather factors, etc., fuzzy logic is used in this embodiment to evaluate vehicle performance, set weight factors, and calculate a comprehensive score:

[0085] a. Input parameters and weight factors:

[0086] 1. Vehicle performance parameters (P i ): P1: maximum load capacity; P2: maximum driving speed; P3: fuel efficiency; P4: maintenance cost; weight factor (w i ): w1=0.3; w2=0.2; w3=0.2; w4=0.1;

[0087] 2. Road condition parameters (Pi ): P5: road type (highway, urban road, rural road, converted into quantitative score); P6: road condition (good, general, poor, converted into quantitative score); weight factor (w i ):w5=0.05;w6=0.05;

[0088] 3. Weather factor parameters (P i ): P7: Current weather conditions (sunny, cloudy, rainy, snowy, foggy, converted into quantitative scores); weight factor (w i ): w7 = 0.1;

[0089] b. Input parameter quantitative scoring table is shown in the following table:

[0090]

[0091] According to the formula S=w1·P1+w2·P2+…+w n ·P n , calculate the adaptability score of each vehicle under the current route and weather conditions. For example, if the quantitative scores of various parameters of the vehicle under the current route and weather conditions are shown in the table above, the adaptability score S is calculated as follows:

[0092] S=0.380+0.290+0.275+0.160+0.0585+0.0590+0.1*70=24+18+15+6+4.25+4.5+7=84.75;

[0093] Based on the suitability score, the vehicle with the highest score is selected and assigned to a specific route. If there are multiple vehicles with similar scores, other factors such as the vehicle's current location and the urgency of the mission are considered to make the final decision.

[0094] (6) Cargo status model construction: Construct a finite element analysis model or a dynamic mechanical model based on real-time monitoring data. For example, the finite element analysis model is used to simulate the cargo reinforcement status.

[0095] (7) State prediction algorithm: Use Kalman filter or particle filter algorithm to predict the future state of the goods. For example, suppose the current state of the goods is X k , predict the state X at the next moment through the Kalman filter algorithm k+1 .

[0096] Formula example:

[0097] X k+1 =F k X k +B k U k +W k

[0098] Among them, F k is the state transition matrix, B k is the control matrix, U k is the control input, W k is the process noise.

[0099] (8) Collision risk assessment algorithm: Use Monte Carlo simulation or fault tree analysis to assess collision risk. For example, Monte Carlo simulation is used to simulate the driving trajectory of a vehicle under different road conditions and calculate the probability of a collision.

[0100] Algorithm: Monte Carlo simulation, formula:

[0101]

[0102] Where: is P 碰撞 The probability of a collision, N is the total number of simulations, f i is the collision function of the i-th simulation. If a collision occurs, I(f i )=1, otherwise I(f i )=0.

[0103] Data parameters:

[0104] Vehicle driving trajectory data: including vehicle speed, acceleration, direction, etc.

[0105] Road condition data: including road conditions, traffic flow, traffic signals, etc.

[0106] Environmental data: including weather conditions, visibility, road wetness, etc.

[0107] (9) Safety hazard identification algorithm: Use the support vector machine (SVM) algorithm to identify safety hazards. For example, analyze historical accident data, train the SVM model, and identify potential safety hazards that may lead to accidents.

[0108] Through the above embodiment, the system comprehensively analyzes the vehicle status, road environment, cargo reinforcement status and personnel operation behavior, accurately judges the potential safety threats during transportation, and provides a basis for early warning and operation guidance.

[0109] Algorithm: Support Vector Machine (SVM), formula:

[0110]

[0111] Where: f(x) is the decision function used to determine whether it is a safety hazard. x is the input feature vector. i is a training sample. i Is the label of the training sample, indicating whether there is a safety hazard. iis the Lagrange multiplier. K is the kernel function used to calculate the similarity between samples. b is the bias term.

[0112] Data parameters:

[0113] Historical accident data: including the environment, time, vehicle status, operating behavior, etc. of the accident.

[0114] Feature vector: may include vehicle speed, acceleration, driving trajectory, environmental conditions, cargo reinforcement status, personnel operating behavior, etc.

[0115] Label: Indicates whether there are safety hazards in historical accident data.

[0116] Imagine a truck transporting a batch of large power transformers weighing 120 tons and measuring 15 meters long, 4 meters wide, and 4.5 meters high from Chengdu to a power station in Chongqing. This solution's intelligent sensing and early warning system fully intervenes to ensure safe and efficient transportation.

[0117] During the data cleaning phase, the system collected 1,000 speed data points from onboard sensors and found five missing and three abnormal data points. To process this data, the system calculated the mean speed of the non-missing data to be 29 km / h and used this mean to fill in the missing data. For abnormal data, the system calculated the median of all valid speed data points and replaced the outliers with this median to ensure the accuracy and consistency of the dataset.

[0118] During the data aggregation phase, the system aggregates the cleaned data into 10-minute groups to analyze the changing trends in the vehicle's driving status. For example, the system may find that within a 10-minute period, the average speed of the vehicle has dropped from 30 km / h to 28 km / h, which may indicate that the vehicle has encountered traffic congestion or deteriorating road conditions. At the same time, the system will also classify and summarize the data according to different road sections, such as calculating the average speed and traffic flow on highways and urban roads respectively. In addition, the system will also classify the data into categories such as vehicle status, road environment, and cargo status for analysis to more comprehensively understand the various factors in the transportation process.

[0119] During the data visualization phase, the system generated various charts and heat maps to visually display the data. For example, the system created a line graph showing vehicle position over time, displaying the vehicle's position at different points in time and reflecting the vehicle's driving trajectory. The system also generated a histogram of vehicle speed distribution, showing the number of vehicles in different speed ranges and analyzing the distribution of vehicle speeds. Furthermore, the system created a heat map of the frequency of road accidents, using different colors on the map to represent the frequency of accidents. Darker colors indicate more frequent accidents, providing drivers with intuitive risk warnings.

[0120] During the path matching algorithm phase, the system uses Hausdorff distance or Frechet distance to calculate the degree of deviation between the vehicle's current location and the pre-set route. For example, the system sets a Hausdorff distance threshold of 0.1 km. If the calculated deviation exceeds this threshold, it is considered a significant deviation and may trigger an alert. This step helps ensure that the vehicle adheres to the pre-set route and reduces the risk of deviation.

[0121] During the comprehensive evaluation algorithm phase, the system considers vehicle performance parameters, road conditions, weather factors, and other factors, using fuzzy logic or Bayesian networks to calculate a suitability score. For example, the system uses fuzzy logic to evaluate vehicle performance, assign weighting factors, and calculate a comprehensive score. This score helps assess the vehicle's performance and safety under current conditions.

[0122] During the cargo status modeling phase, the system uses real-time monitoring data to construct a finite element analysis model or dynamic mechanical model. For example, a finite element analysis model can be used to simulate the cargo reinforcement status. This model helps assess the stability of cargo during transportation and ensure its safe arrival at its destination.

[0123] In the state prediction algorithm stage, the system uses Kalman filter or particle filter algorithm to predict the future state of the goods. For example, if the current state of the goods is X k The system predicts the next state X through the Kalman filter algorithm k+1 .

[0124] During the collision risk assessment algorithm phase, the system uses Monte Carlo simulation or fault tree analysis to assess collision risk. For example, the system simulates vehicle trajectories under different road conditions through Monte Carlo simulation and calculates the probability of a collision. This assessment helps identify high-risk sections of road and provide early warning of risks.

[0125] Finally, during the safety hazard identification algorithm phase, the system utilizes a support vector machine (SVM) algorithm to identify potential safety hazards. For example, the system analyzes historical accident data to train an SVM model and identify potential safety hazards that could lead to accidents. This identification helps the system comprehensively analyze vehicle status, road conditions, cargo reinforcement status, and operator behavior, accurately assessing potential safety threats during transportation and providing a scientific basis for early warning and operational guidance.

[0126] The intelligent sensing and early warning system can comprehensively monitor the transportation process, timely detect and warn potential safety hazards, thereby improving the safety and efficiency of transportation. The application of the system can not only reduce the probability of accidents and transportation costs, but also improve transportation efficiency, providing a solid logistical support for the infrastructure construction of the power industry. The data analysis and risk assessment results are shown as follows Figure 3 As shown; the data analysis and risk assessment steps workflow reference diagram is as follows Figure 4 shown.

[0127] Step 4: Warning and Operational Guidance:

[0128] During the early warning and operational guidance phase, the system's response mechanism is designed to be both rapid and detailed, ensuring prompt and effective action in emergency situations. In addition to alerting drivers and operators through sound, lighting, and display screens, the system also offers a variety of interactive methods to enhance the effectiveness of early warnings.

[0129] Multi-channel warning: In addition to basic sound and light alarms, the system also pushes notifications via SMS, email or mobile applications to ensure that drivers and operators can receive warning information in a timely manner even if they are not in the cab.

[0130] Interactive display: The display in the cab not only displays warning information, but also provides a detailed map view and animated simulation to intuitively show potential safety hazards and their possible impact range, helping personnel better understand and take appropriate measures.

[0131] Voice Command: The system supports voice command function, which can automatically play preset voice warnings or operating instructions, especially suitable for use when the driver's vision is limited or his hands are busy operating.

[0132] The operational guidance module provides personalized risk avoidance advice or operational guidance based on the specific content of the warning. These suggestions and guidance are based on the prediction results of big data analysis and machine learning models, and take into account multiple factors such as current traffic conditions, vehicle status, cargo status, and personnel status.

[0133] Dynamic path planning: When the system finds that the vehicle is about to deviate from the preset route, it not only suggests the driver to slow down or adjust the driving direction, but also dynamically plans a new optimal path based on real-time traffic information and displays it in real time on the cab display screen.

[0134] Cargo reinforcement suggestions: If the cargo reinforcement status is abnormal, the system will give specific reinforcement measures based on the nature, size, weight and stress conditions of the cargo during transportation, such as adding lashing points and adjusting the lashing strength.

[0135] Personnel operation guidance: For illegal operations or unsafe behaviors of drivers and operators, the system not only issues an alarm, but also provides detailed operation guidance videos or graphic descriptions to help personnel correct incorrect behaviors and improve operational standardization.

[0136] The system also possesses self-learning and optimization capabilities. After each warning and operational guidance, the system collects user feedback and actual processing results, and continuously optimizes the warning model and operational guidance strategy through machine learning algorithms to improve the accuracy and effectiveness of the system.

[0137] Step 5: Network information reception and processing:

[0138] As a bridge between the system and the external traffic management system, the network information receiving module undertakes the important task of obtaining and processing the latest traffic information in real time.

[0139] The system not only receives official information from traffic management departments, but also obtains relevant data through interfaces with other traffic information systems (such as road condition monitoring systems and weather forecast systems). These diverse information sources ensure that the system can fully grasp the latest traffic dynamics and weather changes.

[0140] Real-time processing: The system uses efficient data processing algorithms and concurrent processing technologies to ensure that new information can be quickly parsed and processed after it is received, and to update the data models and warning rules within the system in the shortest possible time.

[0141] Dynamic adaptability: Based on updated information, the system can automatically adjust various parameters and strategies during the transportation process, such as adjusting driving speed and selecting alternative routes, to adapt to the ever-changing traffic environment.

[0142] At the same time, the system also has data backup and recovery functions to ensure that data can be quickly restored and normal operation can continue in the event of unexpected situations such as network failure or data loss.

[0143] Step 6: System maintenance and upgrade:

[0144] System maintenance and upgrades are key to ensuring the long-term stability of the system and the continuous improvement of its intelligence level. During this phase, the system takes a variety of measures to maintain the continuous optimization of its performance and functions.

[0145] In summary, this example, through systematic design and refined implementation, achieves comprehensive monitoring and intelligent management of long and large cargo transportation. From system initialization and data input to early warning and operational guidance, network information reception and processing, and system maintenance and upgrades, each link is closely interconnected and mutually supportive, forming an efficient, reliable, and intelligent transportation management system.

[0146] Example 2

[0147] In the case of long-haul freight transportation in the power industry, consider transporting a batch of large power transformers weighing 120 tons from their production site in Chengdu to a power station in Chongqing. This involves complex logistics arrangements and the need to ensure the safety and integrity of the equipment during transportation. Therefore, the application of an intelligent perception and early warning system based on multi-source data fusion for long-haul multimodal freight transportation becomes particularly important in this scenario.

[0148] System initialization and data input:

[0149] Before transportation began, the user entered the transformer's detailed dimensions (15 meters long, 4 meters wide, and 4.5 meters high), precise weight (120 tons), and specific transportation route (starting from the Chengdu manufacturing plant, via the Chengdu-Chongqing Expressway, and finally to the power station in Chongqing) through the system interface. The system received this information in JSON format and verified the accuracy of the data using a built-in validation algorithm (regular expression matching). Next, the system constructed a 3D spatial geometric model based on the input data to describe the device's appearance and, in conjunction with the GIS geographic information system, generated an optimal transportation route.

[0150] Real-time data perception and collection:

[0151] During transportation, the intelligent perception module begins operating. For example, the system uses an integrated high-definition camera to detect a height limit of 4.2 meters on the road ahead. Using image processing algorithms, it determines the distance between the top of the vehicle and the bottom of the bridge is 0.3 meters, close to the preset safety threshold of 0.5 meters. The system then uses a collision warning model and Monte Carlo simulation to predict the probability of collision for vehicles traveling through this section at different speeds. Assuming the current vehicle speed is 25 km / h, the system calculates through simulation that the probability of collision is 5% at this speed, exceeding the preset safety threshold of 3%. The system then displays a warning message on the in-cabin display, accompanied by a voice prompt: "The height limit ahead is too low. Please slow down." The system also notifies transportation management personnel via text message.

[0152] Data analysis and risk assessment:

[0153] After receiving real-time data collected by the perception module, the data analysis module assesses collision risk through Monte Carlo simulation. This algorithm calculates the probability of a collision by simulating the vehicle's trajectory under different road conditions. For example, it simulates a vehicle passing through a height-restricted area at different speeds (20km / h, 25km / h, 30km / h) and different road inclination angles (2 degrees, 3 degrees, 4 degrees), and calculates the corresponding collision probability. Assuming that the collision probability is 5% at a speed of 25km / h, while at a speed of 20km / h, the collision probability is reduced to 2%, the system recommends that the driver slow down to reduce the risk of collision.

[0154] During the real-time data perception and collection phase, the intelligent perception module captured the height limit of the road ahead as 4.2 meters through the integrated high-definition camera. Through the image processing algorithm, the system determined that the distance between the top of the vehicle and the bottom of the bridge was 0.3 meters, close to the preset safety threshold of 0.5 meters. During this process, the system used image processing technology, such as edge detection algorithm (Canny edge detection), to determine the distance between the vehicle and the bottom of the bridge. Assuming that the image resolution captured by the camera is 1920x1080 pixels, the system can use the Canny edge detection algorithm to identify the boundary between the top outline of the vehicle and the bottom of the bridge, and then convert the pixel distance in the image into the actual distance. The key here is the pixel to actual distance conversion factor, which can be calibrated using an object of known size (such as a calibration plate).

[0155] In this embodiment, it is necessary to simulate the situation where a vehicle passes through a height-restricted area at different speeds V and different road inclination angles (θ) to calculate the corresponding collision probability.

[0156] First, define a few key variables:

[0157] Minimum vertical distance Hmin from the top of the vehicle to the ground

[0158] Height limit of the road ahead Hlimit = 4.2m

[0159] Safety threshold Hsafe = 0.5m

[0160] The goal of collision risk assessment is to calculate whether the vertical distance between the top of the vehicle and the bottom of the bridge is greater than or equal to Hsafe at a given speed V and road inclination angle θ, that is:

[0161] Htop ≥ Hlimit - Hsafe

[0162] Where Htop is the vertical distance between the top of the vehicle and the ground. On straight roads, Htop can be directly obtained from the vehicle's design parameters; however, on inclined roads, the change in the vehicle's center of gravity needs to be considered. The following formula can be used to estimate the change in Htop at different inclination angles:

[0163] Htop(θ)=Hmin+Hadjust(θ)

[0164] Hadjust(θ)=Hmin·sin(θ)

[0165] Hadjust(θ) represents the height change due to the tilt angle.

[0166] Improved Algorithm: To improve the algorithm and make it more accurately reflect the actual situation, a vehicle dynamic model and an uncertainty distribution can be introduced. The vehicle dynamic model takes into account the impact of factors such as vehicle acceleration and braking on height, while the uncertainty distribution is the collision probability distribution under different speeds and tilt angles obtained based on historical data statistics.

[0167] In the Monte Carlo simulation, a large number of samples are randomly selected, and each sample represents a combination of a set of speed and tilt angle. For each set of samples, Htop(θ) is calculated and compared with Hlimit - Hsafe. If Htop(θ) < Hlimit - Hsafe, it is considered that a collision will occur. By repeating this process multiple times, the collision probability under given speeds and tilt angles can be estimated.

[0168] For example, simulate the situation where a vehicle passes through a height - restricted area at different speeds (20 km / h, 25 km / h, 30 km / h) and different road tilt angles (2 degrees, 3 degrees, 4 degrees). Assume that at a speed of 20 km / h, the collision probability is 5%, and at a speed of 25 km / h, the collision probability drops to 2%. Therefore, the system advises the driver to slow down to reduce the collision risk.

[0169] Warning and Operation Guidance:

[0170] In the warning and operation guidance stage, the system not only issues alerts to the driver through sound, light, and the display screen, but also pushes notifications via text messages, emails, or mobile applications to ensure that even if the driver and operator are not in the cab, they can receive the warning information in a timely manner. In addition, the system provides a detailed map view and animation simulation to visually display potential safety hazards and their possible impact ranges, helping personnel better understand and take corresponding measures. For example, in the above scenario, the system shows the height limit of the upcoming section on the display screen in the cockpit and simulates the situation of the vehicle passing through this section at different speeds, enabling the driver to clearly understand the current dangerous situation and take deceleration measures according to the system's advice.

[0171] The above - described embodiments only represent the specific implementation manners of the present invention. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. An intelligent perception and active early warning system for safety risks of long and large cargo multimodal transport based on multi-source data fusion, characterized in that: include: Raw data input module: used to receive detailed dimensions, weight, transportation route, personnel information, and transportation tool information of large and long goods input by users, and to establish an initial data model to provide a basis for intelligent analysis; and to import historical transportation data and external data sources; Business scenario intelligent perception module: integrated with high-definition camera, Beidou positioning, inclination sensor, acceleration monitor, tension monitor, tire pressure monitor, vehicle level monitor, and limit laser rangefinder, used to perceive the operation status of transportation tools, cargo fixation status, personnel operation behavior and road environment dynamic information in real time; and used to connect with other intelligent devices to obtain transportation environment information; Data collection and analysis module: Use big data processing technology and artificial intelligence algorithms to conduct in-depth analysis of the data collected by the business scenario intelligent perception module; including the matching degree of the transportation tool's location, speed, direction and preset route; the adaptability of the vehicle's current condition to the road environment; the stability of the cargo reinforcement state; the compliance of personnel operation behavior; and continuously optimize the analysis model and early warning strategy based on historical data and real-time feedback; Safety risk warning module: used to send out warnings to the on-site commander and driver when potential safety hazards are discovered; And for remote monitoring and emergency communications, to contact relevant departments and personnel in time in emergency situations; Operation guidance module: used to provide operation guidance to the transport site commander and driver in combination with warning information; the guidance includes specific emergency measures and operation steps, as well as explanations of potential risks and instructions for preventive measures; it has a voice prompt function to provide timely guidance when the driver's vision is obstructed or his attention is distracted; Networked information receiving module: It is connected to the external system to receive and process the latest traffic information and weather information in real time. According to the received information, it can automatically adjust the early warning strategy and operation guidance plan to deal with emergencies and complex environments.

2. A method for intelligent perception and active early warning of safety risks of long and large cargo multimodal transport based on multi-source data fusion, characterized in that: The method is implemented by the intelligent perception and active early warning system for safety risks of multimodal transport of long and large cargo based on multi-source data fusion as claimed in claim 1, wherein the method comprises the following steps: Step 1: System initialization and data input: After logging into the system, the user first fills in the detailed dimensions, precise weight and specific transportation route information of the long and large goods through the user interface. Then, the input information is automatically verified using the built-in verification algorithm to ensure the integrity and rationality of the data; Step 2: Real-time data perception and collection: During the transportation process, the business scenario intelligent perception module continuously perceives the vehicle operation status in real time; Step 3: Data analysis and risk assessment: After the data analysis module receives the real-time data collected by the perception module, it uses big data processing technology and artificial intelligence algorithms to conduct in-depth analysis; Step 4: Warning and operation guidance: Through multi-channel warning, ensure that the driver and operator can receive warning information in time even if they are not in the cab; Step 5: Network information reception and processing: Receive official information from the traffic management department and obtain relevant data through interfaces with other traffic information systems; Step 6: System maintenance and upgrade: Take a variety of measures to maintain continuous optimization of its performance and functions.

3. The method for intelligent perception and active early warning of safety risks of long and large cargo multimodal transport based on multi-source data fusion according to claim 2 is characterized in that: The step 1 is specifically as follows: The system builds an initial data model based on the verified data, including the description of the equipment shape and size model through a three-dimensional space geometry algorithm, and the generation of an optimal or feasible path transportation route model based on a GIS geographic information system and a path planning algorithm; at the same time, the system automatically connects and initializes various sensor devices through a built-in communication protocol, including vehicle operation status sensors, environmental perception sensors, cargo reinforcement status monitoring sensors, and personnel operation behavior monitoring sensors.

4. The method for intelligent perception and active early warning of safety risks of long and large cargo multimodal transport based on multi-source data fusion according to claim 2 is characterized in that: In step 2, for the monitoring of key vehicle parameters, the sensor data acquisition algorithm is used to ensure the accuracy and real-time nature of the data; Beidou positioning information is obtained according to the satellite positioning algorithm to calculate the vehicle position, speed, and direction information in real time; High-definition cameras and radars use image processing algorithms and radar detection algorithms to capture real-time information about the surrounding environment, providing the system with data on road width, height limit, and turning angle. By comparing the vehicle's operating status with road environment information, the adaptability assessment algorithm is used to assess the adaptability of the vehicle to the environment in real time, ensuring a safe and smooth transportation process. When monitoring the reinforcement status of cargo, the lashing tension monitoring and impact vibration detectors use mechanical models and signal processing algorithms to monitor the stress and vibration status of the cargo in real time, and immediately trigger the early warning mechanism when any abnormality is found; at the same time, face recognition monitoring, fatigue recognition, and distraction recognition equipment continuously monitor personnel's operating behavior through deep learning algorithms to ensure that the driver and operator are in good condition.

5. The method for intelligent perception and active early warning of safety risks of long and large cargo multimodal transport based on multi-source data fusion according to claim 4 is characterized in that: In step 3, big data processing technologies include data cleaning, data aggregation, and data visualization; artificial intelligence algorithms include machine learning algorithms and deep learning algorithms; in terms of the matching analysis of vehicle position, speed, direction and preset route, a path matching algorithm is used to evaluate the degree of deviation between the vehicle's current position and the preset route, and adjustment suggestions are given in combination with road condition information; the adaptability assessment of the vehicle's condition and road environment is performed through a comprehensive evaluation algorithm, using fuzzy logic or Bayesian network methods to calculate the adaptability score. When evaluating the stability of the cargo reinforcement status, a cargo status model is built through real-time monitoring data, and the status prediction algorithm is used to predict the future status of the cargo. When an abnormality is found, an early warning mechanism is triggered. The compliance assessment of personnel operation behavior is based on the behavior recognition results of the deep learning algorithm, combined with the operation specification library to make compliance judgments. When an illegal operation is found, an alarm is immediately issued and operational guidance suggestions are provided. Using the collision risk assessment algorithm, we comprehensively consider multiple factors such as vehicle status, road environment, cargo reinforcement status, and personnel operating behavior to accurately determine whether there are potential safety threats in the current transportation process, and provide important basis for subsequent early warning and operational guidance.

6. The method for intelligent perception and active early warning of safety risks of long and large cargo multimodal transport based on multi-source data fusion according to claim 2 is characterized in that: In step 4, multi-channel warning includes not only sound and light alarms, but also SMS, email or mobile application push notifications; as well as: Interactive display: The display in the cab is used not only to display alarm information, but also to provide detailed map views and animation simulations to intuitively display potential safety hazards and their possible impact range, helping personnel to better understand and take corresponding measures; Voice command: The system supports voice command function, which automatically plays preset voice warnings or operating instructions for use when the driver's vision is limited or his hands are busy operating; The operation guidance module is used to provide personalized risk avoidance suggestions or operation guidance according to the specific content of the warning, based on the prediction results of big data analysis and machine learning models, and taking into account the current traffic conditions, vehicle status, cargo conditions and personnel status factors; Dynamic route planning: When the vehicle is about to deviate from the preset route, the system recommends the driver to slow down or adjust the driving direction, and dynamically plans a new optimal route based on real-time traffic information, which is displayed in real time on the display screen in the cab; Cargo reinforcement suggestions: For abnormal cargo reinforcement status, specific reinforcement measures are given according to the nature, size, weight and stress of the cargo during transportation; Personnel operation guidance: Alarms are issued for illegal operations or unsafe behaviors of drivers and operators, and operation guidance videos or graphic descriptions are provided to help personnel correct wrong behaviors and improve operational standardization; After each warning and operation guidance, user feedback information and actual processing results are collected, and the warning model and operation guidance strategy are optimized through machine learning algorithms to improve the accuracy and effectiveness of the system.

Images