AI intelligent auxiliary device for road tunnel engineering construction
By using multimodal fusion perception module, deep learning and reinforcement learning algorithm, edge computing and distributed AI in the AI intelligent auxiliary devices for road tunnel engineering construction, the problems of limited perception capabilities and decision-making delays in complex environments are solved, and more efficient and safer tunnel construction assistance is achieved.
Patent Information
- Application Number
- CN202411918113.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-25
- Publication Date
- 2025-04-18
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing AI intelligent auxiliary devices for road tunnel construction have limited perception capabilities in complex environments, lack independent learning and decision-making capabilities, rely on central servers for data processing, and have latency and bandwidth bottlenecks, which are difficult to meet real-time requirements.
The multimodal fusion perception module is adopted, combined with deep learning and reinforcement learning algorithms, edge computing and distributed AI are realized, AI-driven risk management and human-computer interaction functions are enhanced, and perception accuracy and decision-making speed are improved through the fusion of multiple sensor data and local data processing.
It realizes more comprehensive and accurate environmental perception, improves the intelligence level and adaptability of AI devices, reduces data transmission delay and bandwidth requirements, and enhances construction safety and efficiency.
Smart Images
Figure BDA0005207285610000031 
Figure BDA0005207285610000081 
Figure FDA0005207285600000021
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent assistance technology for tunnel engineering, and specifically to an AI intelligent assistance device for the construction of road tunnel engineering. Background Art
[0002] Tunnel engineering is a structure built underground, underwater or in the mountains to lay railways or build roads for motor vehicles to pass. According to its location, it can be divided into three categories: mountain tunnel engineering, underwater tunnel engineering and urban tunnel engineering. The tunnel construction environment is complex, and there are risks such as collapse, gas explosion, and fire. The manual monitoring and early warning capabilities are limited. In recent years, AI technology has gradually been introduced into tunnel engineering construction, which can realize real-time perception of the tunnel construction environment, intelligent decision-making and risk warning, thereby improving construction safety and efficiency.
[0003] There are still some problems in the use of existing AI intelligent auxiliary devices for road tunnel engineering construction. Traditional AI devices mainly rely on a single type of sensor and have limited perception capabilities in complex environments. AI devices lack autonomous learning and decision-making capabilities, and are difficult to adapt to complex construction environments and needs. They cannot fully grasp the environmental information in the tunnel and are difficult to conduct effective risk warnings and decision support. Traditional AI devices rely on central servers for data processing, which has delays and bandwidth bottlenecks and is difficult to meet real-time requirements. Therefore, technical personnel in this field provide an AI intelligent auxiliary device for road tunnel engineering construction to solve the problems raised in the above background technology. Summary of the invention
[0004] 1. Technical issues to be resolved
[0005] In view of the shortcomings of the prior art, the present invention provides an AI intelligent auxiliary device for road tunnel engineering construction, which solves the problems that traditional AI devices mainly rely on a single type of sensor, have limited perception capabilities in complex environments, lack autonomous learning and decision-making capabilities, and are difficult to adapt to complex construction environments and needs; traditional AI devices rely on central servers for data processing, have delays and bandwidth bottlenecks, and are difficult to meet real-time requirements.
[0006] (II) Technical solution
[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: an AI intelligent auxiliary device for road tunnel engineering construction, comprising:
[0008] Multimodal fusion perception module, deep learning and reinforcement learning algorithms, edge computing and distributed AI, AI-driven risk management, human-computer interaction module, power supply and energy management module, mechanical structure and protection module;
[0009] The multimodal fusion perception module also includes a sensor fusion module and a data fusion and processing module. The sensor fusion module also includes a visual sensor, a laser radar, an infrared thermal imager, an acoustic wave sensor, a gas sensor and a humidity sensor;
[0010] The data fusion and processing module is used to integrate data from different sensors and perform preprocessing and analysis to improve the accuracy and reliability of environmental perception;
[0011] The deep learning and reinforcement learning algorithms also include an AI computing module and a training and optimization module. The AI computing module also includes a high-performance AI processor, a deep learning algorithm library, and a reinforcement learning algorithm library.
[0012] The training and optimization module is used for model training and optimization to ensure that the AI algorithm can adapt to different construction environments and requirements;
[0013] The edge computing and distributed AI also include an edge computing module and a distributed AI system. The edge computing module realizes local data processing and real-time decision-making, and the edge AI chip is used for efficient data processing and low-latency decision-making;
[0014] The distributed computing nodes distribute AI computing tasks to multiple nodes to achieve efficient data processing and decision-making;
[0015] The AI-driven risk management also includes a risk prediction and analysis module and an early warning and decision support module. The risk prediction and analysis module uses AI algorithms to predict and analyze risks and identify potential risk points in advance. The data collection and analysis module collects environmental data and construction data in the tunnel in real time to perform dynamic risk assessment.
[0016] The early warning and decision support module provides real-time early warning and decision support to help construction management personnel adjust construction plans and response strategies in a timely manner.
[0017] Preferably, the human-computer interaction module includes a touch screen display and a speech recognition and synthesis module for interaction between the user and the system; the power and energy management module includes a battery pack and a power management system to ensure stable power supply and energy optimization of the device; the mechanical structure and protection module includes a sturdy and durable casing and a waterproof and dustproof design to ensure the normal operation of the device in harsh environments.
[0018] Preferably, the multimodal fusion perception module can obtain more comprehensive and accurate environmental information by fusing multiple sensor data. The detailed steps are as follows:
[0019] A1. Data collection: using a variety of sensors, including cameras, lidar, infrared thermal imagers, acoustic wave sensors, gas sensors and humidity sensors to collect real-time environmental data in the tunnel;
[0020] A2. Data preprocessing: cleaning, denoising and normalizing the collected data to improve data quality;
[0021] A3. Feature extraction: extract features from different sensor data, extract edge, texture and other features from image data, and extract distance and shape features from lidar data;
[0022] A4. Data fusion: Use the data fusion algorithm Kalman filter to fuse data from different sensors to form a comprehensive perception of the tunnel environment. The algorithm formula is as follows:
[0023]
[0024] P k =(IK k H)P k-1
[0025] in, is the state estimate, Z k is the measurement value, H is the observation matrix, K k is the Kalman gain, P k is the error cosquare matrix.
[0026] Preferably, the deep learning and reinforcement learning algorithms can learn complex image recognition and behavior prediction, and the detailed steps to improve the intelligence level are as follows:
[0027] B1. Data collection: collect a large amount of data related to tunnel construction, including image data, sensor data and construction logs;
[0028] B2. Data annotation: annotate image data and mark different objects and scenes;
[0029] B3. Model training: Use deep learning algorithms to train models, including convolutional neural networks and recurrent neural networks, to learn image recognition and behavior prediction. The algorithm formula is as follows:
[0030] Convolutional Neural Networks:
[0031] y=f(W·x+b)
[0032] Where y is the output, x is the input, W is the weight matrix, b is the bias, and f is the activation function;
[0033] Recurrent Neural Networks:
[0034] h t=tanh(W hx x t +W hh h t-1 +b h )
[0035] y t =W yh h t +b h
[0036] Among them, h t is the hidden state, x t is the input, y t is the output, W is the weight matrix, and b is the bias
[0037] B4. Strategy learning: Use reinforcement learning algorithms to learn strategies and optimize construction paths and equipment control.
[0038] Preferably, the edge computing and distributed AI real-time data processing and decision-making to improve the response speed of the system are as follows:
[0039] C1. Edge computing: pre-processing data on AI devices to reduce data transmission volume, and performing real-time data processing and decision-making on AI devices to improve the response speed of the system;
[0040] C2. Distributed AI distributes AI computing tasks to multiple nodes to achieve efficient data processing and decision-making, realize the collaborative work of multiple AI devices, and improve overall efficiency and security.
[0041] Preferably, the AI-driven risk management identifies, evaluates, predicts and warns various risks in the tunnel construction process, thereby improving construction safety, reducing accident rates and optimizing construction plans. The detailed steps are as follows:
[0042] S1. Data collection: collect multi-source data related to construction risks, including temperature, humidity, gas concentration, vibration and noise, construction personnel behavior captured by video data monitoring cameras, equipment operation status and environmental changes, construction logs, construction progress, equipment maintenance records and personnel scheduling, historical accident data, accident records in similar projects in the past, cause analysis, etc.
[0043] S2. Data preprocessing, cleaning, removing noise data, filling missing values, normalization, normalizing different types of data to make them have the same dimension, feature extraction, extracting risk-related features from raw data, extracting personnel behavior features from video data, such as whether they wear a helmet, extracting environmental anomaly features from sensor data, and extracting construction progress, equipment status and other features from construction logs;
[0044] S3. Data training: Use machine learning algorithms to train the collected data, identify potential risk factors, and classify risks;
[0045] S4. Risk prediction and assessment: Use time series prediction models to predict risks and assess the likelihood and impact of risks;
[0046] S5. Risk warning and decision support. Based on the risk identification and prediction results, the AI system generates risk warning information and notifies relevant personnel through a visual interface and alarm system. The AI system provides decision support to help construction managers develop risk response strategies and optimize construction plans.
[0047] Preferably, the risk warning and decision support also include a warning mechanism and decision support. The warning mechanism sets a risk threshold. When the risk prediction value exceeds the threshold, the warning is automatically triggered. According to the risk type and level, corresponding warning information and processing suggestions are provided. The decision support provides a risk assessment report, including the probability of risk occurrence, the degree of impact and possible consequences, and provides risk response suggestions, adjusts the construction progress, changes the construction methods and strengthens safety measures.
[0048] (III) Beneficial effects
[0049] The present invention provides an AI intelligent auxiliary device for road tunnel engineering construction. It has the following beneficial effects:
[0050] 1. In the present invention, by fusing multiple sensor data, more comprehensive and accurate environmental information can be obtained, and the perception accuracy can be improved. Multimodal fusion can make up for the shortcomings of a single sensor under certain environmental conditions, improve the robustness of the system, and be able to perceive information that a single sensor cannot capture.
[0051] 2. In the present invention, the deep learning algorithm can learn complex image recognition and behavior prediction, improve the intelligence level of AI devices, the reinforcement learning algorithm can realize autonomous learning and decision-making, optimize construction paths and equipment control, and the AI algorithm can adapt to different construction environments and needs through continuous learning.
[0052] 3. In the present invention, edge computing can realize local data processing and real-time decision-making, reduce data transmission delay, reduce data transmission volume, and save bandwidth resources. Edge computing can realize real-time data processing and decision-making, improve the response speed of the system, and the distributed AI system can expand computing nodes as needed to improve the processing power of the system. The distributed AI system has higher fault tolerance, and the failure of a single node will not affect the operation of the entire system.
[0053] 4. In the present invention, through real-time monitoring and risk warning, potential safety hazards can be discovered and eliminated in time, the accident rate can be reduced, and the construction progress and methods can be adjusted according to the risk prediction results, the construction plan can be optimized, the construction efficiency can be improved, the risks can be identified and evaluated in advance, and the economic losses caused by the occurrence of risks can be reduced. The data support and decision-making suggestions provided by the AI system can help construction managers make more scientific decisions. DETAILED DESCRIPTION
[0054] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0055] Embodiment 1:
[0056] An embodiment of the present invention provides an AI intelligent auxiliary device for road tunnel engineering construction, including:
[0057] Multimodal fusion perception module, deep learning and reinforcement learning algorithms, edge computing and distributed AI, AI-driven risk management, human-computer interaction module, power supply and energy management module, mechanical structure and protection module;
[0058] The multimodal fusion perception module also includes a sensor fusion module and a data fusion and processing module. The sensor fusion module also includes a visual sensor, a lidar, an infrared thermal imager, an acoustic sensor, a gas sensor and a humidity sensor.
[0059] The data fusion and processing module is used to integrate data from different sensors and perform preprocessing and analysis to improve the accuracy and reliability of environmental perception;
[0060] Deep learning and reinforcement learning algorithms also include AI computing modules and training and optimization modules. The AI computing module also includes a high-performance AI processor, a deep learning algorithm library, and a reinforcement learning algorithm library.
[0061] The training and optimization module is used for model training and optimization to ensure that the AI algorithm can adapt to different construction environments and requirements;
[0062] Edge computing and distributed AI also include edge computing modules and distributed AI systems. Edge computing modules enable local data processing and real-time decision-making, and edge AI chips are used for efficient data processing and low-latency decision-making.
[0063] Distributed computing nodes distribute AI computing tasks to multiple nodes to achieve efficient data processing and decision-making;
[0064] AI-driven risk management also includes risk prediction and analysis modules and early warning and decision support modules. The risk prediction and analysis module uses AI algorithms to predict and analyze risks and identify potential risk points in advance. The data collection and analysis module collects environmental data and construction data in the tunnel in real time to conduct dynamic risk assessment.
[0065] The early warning and decision support module provides real-time early warning and decision support to help construction managers adjust construction plans and response strategies in a timely manner.
[0066] The human-computer interaction module includes a touch screen display and a speech recognition and synthesis module for interaction between the user and the system. The power and energy management module includes a battery pack and a power management system to ensure stable power supply and energy optimization of the device. The mechanical structure and protection module includes a sturdy and durable casing and a waterproof and dustproof design to ensure the normal operation of the device in harsh environments.
[0067] Preferably, the multimodal fusion perception module can obtain more comprehensive and accurate environmental information by fusing multiple sensor data. The detailed steps are as follows:
[0068] A1. Data collection: using a variety of sensors, including cameras, lidar, infrared thermal imagers, acoustic wave sensors, gas sensors and humidity sensors to collect real-time environmental data in the tunnel;
[0069] A2. Data preprocessing: cleaning, denoising and normalizing the collected data to improve data quality;
[0070] A3. Feature extraction: extract features from different sensor data, extract edge, texture and other features from image data, and extract distance and shape features from lidar data;
[0071] A4. Data fusion: Use the data fusion algorithm Kalman filter to fuse data from different sensors to form a comprehensive perception of the tunnel environment. The algorithm formula is as follows:
[0072]
[0073] P k =(IK k H)P k-1
[0074] in, is the state estimate, Z k is the measurement value, H is the observation matrix, K k is the Kalman gain, P kIt is the error co-square matrix. By fusing data from multiple sensors, more comprehensive and accurate environmental information can be obtained and the perception accuracy can be improved. Multimodal fusion can make up for the shortcomings of a single sensor under certain environmental conditions, improve the robustness of the system, and perceive information that a single sensor cannot capture.
[0075] Deep learning and reinforcement learning algorithms can learn complex image recognition and behavior prediction. The detailed steps to improve the level of intelligence are as follows:
[0076] B1. Data collection: collect a large amount of data related to tunnel construction, including image data, sensor data and construction logs;
[0077] B2. Data annotation: annotate image data and mark different objects and scenes;
[0078] B3. Model training: Use deep learning algorithms to train models, including convolutional neural networks and recurrent neural networks, to learn image recognition and behavior prediction. The algorithm formula is as follows:
[0079] Convolutional Neural Networks:
[0080] y=f(W·x+b)
[0081] Where y is the output, x is the input, W is the weight matrix, b is the bias, and f is the activation function;
[0082] Recurrent Neural Networks:
[0083] h t =tanh(W hx x t +W hh h t-1 +b h )
[0084] y t =W yh h t +b h
[0085] Among them, h t is the hidden state, x t is the input, y t is the output, W is the weight matrix, and b is the bias
[0086] B4. Strategy learning: Use reinforcement learning algorithms for strategy learning to optimize construction paths and equipment control. Deep learning algorithms can learn complex image recognition and behavior prediction to improve the intelligence level of AI devices. Reinforcement learning algorithms can achieve autonomous learning and decision-making, optimize construction paths and equipment control, and AI algorithms can adapt to different construction environments and needs through continuous learning.
[0087] Edge computing and distributed AI real-time data processing and decision-making to improve the system's response speed. The detailed steps are as follows:
[0088] C1. Edge computing: pre-processing data on AI devices to reduce data transmission volume, and performing real-time data processing and decision-making on AI devices to improve the response speed of the system;
[0089] C2. Distributed AI distributes AI computing tasks to multiple nodes to achieve efficient data processing and decision-making, realize the collaborative work of multiple AI devices, and improve overall efficiency and security. Edge computing can realize local data processing and real-time decision-making, reduce data transmission delays, reduce data transmission volume, and save bandwidth resources. Edge computing can realize real-time data processing and decision-making, and improve the response speed of the system. Distributed AI systems can expand computing nodes as needed to improve the processing power of the system. Distributed AI systems have higher fault tolerance, and the failure of a single node will not affect the operation of the entire system.
[0090] AI-driven risk management identifies, evaluates, predicts and warns of various risks during tunnel construction, thereby improving construction safety, reducing accident rates and optimizing construction plans. The detailed steps are as follows:
[0091] S1. Data collection: collect multi-source data related to construction risks, including temperature, humidity, gas concentration, vibration and noise, construction personnel behavior captured by video data monitoring cameras, equipment operation status and environmental changes, construction logs, construction progress, equipment maintenance records and personnel scheduling, historical accident data, accident records in similar projects in the past, cause analysis, etc.
[0092] S2. Data preprocessing, cleaning, removing noise data, filling missing values, normalization, normalizing different types of data to make them have the same dimension, feature extraction, extracting risk-related features from raw data, extracting personnel behavior features from video data, such as whether they wear a helmet, extracting environmental anomaly features from sensor data, and extracting construction progress, equipment status and other features from construction logs;
[0093] S3. Data training: Use machine learning algorithms to train the collected data, identify potential risk factors, and classify risks;
[0094] S4. Risk prediction and assessment: Use time series prediction models to predict risks and assess the likelihood and impact of risks;
[0095] S5. Risk warning and decision support. Based on the risk identification and prediction results, the AI system generates risk warning information and notifies relevant personnel through a visual interface and alarm system. The AI system provides decision support to help construction managers develop risk response strategies and optimize construction plans. Through real-time monitoring and risk warning, safety hazards can be discovered and eliminated in a timely manner to reduce the accident rate. Based on the risk prediction results, the construction progress and methods can be adjusted, the construction plan can be optimized, construction efficiency can be improved, risks can be identified and assessed in advance, and economic losses caused by risks can be reduced. The data support and decision-making suggestions provided by the AI system help construction managers make more scientific decisions.
[0096] Risk warning and decision support also include warning mechanism and decision support. The warning mechanism sets a risk threshold. When the risk prediction value exceeds the threshold, the warning is automatically triggered. According to the risk type and level, corresponding warning information and handling suggestions are provided. Decision support provides risk assessment reports, including the probability of risk occurrence, the degree of impact and possible consequences, and provides risk response suggestions, adjusts the construction progress, changes the construction methods and strengthens safety measures.
[0097] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. An AI intelligent auxiliary device for road tunnel engineering construction, characterized in that: include: Multimodal fusion perception module, deep learning and reinforcement learning algorithms, edge computing and distributed AI, AI-driven risk management, human-computer interaction module, power supply and energy management module, mechanical structure and protection module; The multimodal fusion perception module also includes a sensor fusion module and a data fusion and processing module. The sensor fusion module also includes a visual sensor, a laser radar, an infrared thermal imager, an acoustic wave sensor, a gas sensor and a humidity sensor; The data fusion and processing module is used to integrate data from different sensors and perform preprocessing and analysis to improve the accuracy and reliability of environmental perception; The deep learning and reinforcement learning algorithms also include an AI computing module and a training and optimization module. The AI computing module also includes a high-performance AI processor, a deep learning algorithm library, and a reinforcement learning algorithm library. The training and optimization module is used for model training and optimization to ensure that the AI algorithm can adapt to different construction environments and requirements; The edge computing and distributed AI also include an edge computing module and a distributed AI system. The edge computing module realizes local data processing and real-time decision-making, and the edge AI chip is used for efficient data processing and low-latency decision-making; The distributed computing nodes distribute AI computing tasks to multiple nodes to achieve efficient data processing and decision-making; The AI-driven risk management also includes a risk prediction and analysis module and an early warning and decision support module. The risk prediction and analysis module uses AI algorithms to predict and analyze risks and identify potential risk points in advance. The data collection and analysis module collects environmental data and construction data in the tunnel in real time to perform dynamic risk assessment. The early warning and decision support module provides real-time early warning and decision support to help construction management personnel adjust construction plans and response strategies in a timely manner.
2. The AI intelligent auxiliary device for road tunnel engineering construction according to claim 1 is characterized in that: The human-computer interaction module includes a touch screen display and a speech recognition and synthesis module for interaction between the user and the system. The power and energy management module includes a battery pack and a power management system to ensure stable power supply and energy optimization of the device. The mechanical structure and protection module includes a sturdy and durable shell and a waterproof and dustproof design to ensure the normal operation of the device in harsh environments.
3. The AI intelligent auxiliary device for road tunnel engineering construction according to claim 1 is characterized in that: The multimodal fusion perception module can obtain more comprehensive and accurate environmental information by fusing multiple sensor data. The detailed steps are as follows: A1. Data collection: using a variety of sensors, including cameras, lidar, infrared thermal imagers, acoustic wave sensors, gas sensors and humidity sensors to collect real-time environmental data in the tunnel; A2. Data preprocessing: cleaning, denoising and normalizing the collected data to improve data quality; A3. Feature extraction: extract features from different sensor data, extract edge, texture and other features from image data, and extract distance and shape features from lidar data; A4. Data fusion: Use the data fusion algorithm Kalman filter to fuse data from different sensors to form a comprehensive perception of the tunnel environment. The algorithm formula is as follows: P k = ( I-K k H ) P k-1 in, is the state estimate, Z k is the measurement value, H is the observation matrix, K k is the Kalman gain, P k is the error cosquare matrix.
4. The AI intelligent auxiliary device for road tunnel engineering construction according to claim 1 is characterized in that: The deep learning and reinforcement learning algorithms can learn complex image recognition and behavior prediction. The detailed steps to improve the intelligence level are as follows: B1. Data collection: collect a large amount of data related to tunnel construction, including image data, sensor data and construction logs; B2. Data annotation: annotate image data and mark different objects and scenes; B3. Model training: Use deep learning algorithms to train models, including convolutional neural networks and recurrent neural networks, to learn image recognition and behavior prediction. The algorithm formula is as follows: Convolutional Neural Networks: y=f(W·x+b) Where y is the output, x is the input, W is the weight matrix, b is the bias, and f is the activation function; Recurrent Neural Networks: h t =tanh(W hx x t +W hh h t-1 +b h ) y t =W yh h t +b h Among them, h t is the hidden state, x t is the input, y t is the output, W is the weight matrix, and b is the bias B4. Strategy learning: Use reinforcement learning algorithms to learn strategies and optimize construction paths and equipment control.
5. The AI intelligent auxiliary device for road tunnel engineering construction according to claim 1 is characterized in that: The detailed steps of edge computing and distributed AI real-time data processing and decision-making to improve the system response speed are as follows: C1. Edge computing: pre-processing data on AI devices to reduce data transmission volume, and performing real-time data processing and decision-making on AI devices to improve the response speed of the system; C2. Distributed AI distributes AI computing tasks to multiple nodes to achieve efficient data processing and decision-making, realize the collaborative work of multiple AI devices, and improve overall efficiency and security.
6. The AI intelligent auxiliary device for road tunnel engineering construction according to claim 1 is characterized in that: The AI-driven risk management identifies, evaluates, predicts and warns various risks in the tunnel construction process, thereby improving construction safety, reducing accident rates and optimizing construction plans. The detailed steps are as follows: S1. Data collection: collect multi-source data related to construction risks, including temperature, humidity, gas concentration, vibration and noise, construction personnel behavior captured by video data monitoring cameras, equipment operation status and environmental changes, construction logs, construction progress, equipment maintenance records and personnel scheduling, historical accident data, accident records in similar projects in the past, cause analysis, etc. S2. Data preprocessing, cleaning, removing noise data, filling missing values, normalization, normalizing different types of data to make them have the same dimension, feature extraction, extracting risk-related features from raw data, extracting personnel behavior features from video data, such as whether they wear a helmet, extracting environmental anomaly features from sensor data, and extracting construction progress, equipment status and other features from construction logs; S3. Data training: Use machine learning algorithms to train the collected data, identify potential risk factors, and classify risks; S4. Risk prediction and assessment: Use time series prediction models to predict risks and assess the likelihood and impact of risks; S5. Risk warning and decision support. Based on the risk identification and prediction results, the AI system generates risk warning information and notifies relevant personnel through a visual interface and alarm system. The AI system provides decision support to help construction managers develop risk response strategies and optimize construction plans.
7. The AI intelligent auxiliary device for road tunnel engineering construction according to claim 6 is characterized by: The risk warning and decision support also include warning mechanism and decision support. The warning mechanism sets a risk threshold. When the risk prediction value exceeds the threshold, the warning is automatically triggered. According to the risk type and level, corresponding warning information and processing suggestions are provided. The decision support provides a risk assessment report, including the probability of risk occurrence, the degree of impact and possible consequences, and provides risk response suggestions, adjusts the construction progress, changes the construction methods and strengthens safety measures.