CN-121973177-A - Underground cavity disease body detection robot control system

CN121973177ACN 121973177 ACN121973177 ACN 121973177ACN-121973177-A

Abstract

The invention relates to the technical field of underground detection, in particular to a robot control system for detecting an underground cavity disease body, the system comprises an underground data acquisition module, a geological structure analysis module, a cavity and crack analysis module, an underground risk detection module, a geological risk dynamic evaluation module, an emergency path planning module, a detection route optimization module and an underground robot cooperation strategy module. According to the invention, the accuracy and efficiency of underground data acquisition are improved by combining a sensor network technology and machine learning preprocessing, the recognition capability of holes and cracks is improved by using a geological feature analysis and crack recognition technology, the recognition and analysis of potential risks are enhanced by combining a risk assessment algorithm and environmental impact analysis, the application of a dynamic planning algorithm and a risk prediction model is realized, the dynamic assessment and trend analysis of geological risks, the combination of emergency path planning and detection route optimization, the detection efficiency is optimized, and the task allocation and cooperation path planning are optimized by using a multi-robot coordinated control algorithm and a communication synchronization technology.

Inventors

XIA QINGYONG
Xue Dianshun
ZHANG JIANMIN
LIANG ZHIHUA

Assignees

天津蓝图工程技术有限公司
北京数字朗博科技有限公司

Dates

Publication Date: 20260505
Application Date: 20240312

Claims (10)

1. The system is characterized by comprising an underground data acquisition module, a geological structure analysis module, a cavity and crack analysis module, an underground risk detection module, a geological risk dynamic evaluation module, an emergency path planning module, a detection route optimization module and an underground robot cooperation strategy module; The underground data acquisition module is based on an underground detection environment, adopts a sensor network technology and a machine learning preprocessing method, collects stratum data and temperature and humidity information, performs data filtering and feature extraction, performs normalization processing on the data, and generates cleaned and standardized underground data; the geological structure analysis module is used for creating a three-dimensional model of a subsurface stratum based on the cleaned and standardized underground data by adopting a geological modeling algorithm and a three-dimensional visualization technology, analyzing the distribution and composition characteristics of the stratum, and generating a stratum structure three-dimensional model; The cavity and crack analysis module is used for identifying underground cavities and cracks based on a stratum structure three-dimensional model by adopting a geological feature analysis method and a crack identification technology, evaluating the sizes and the shapes of the underground cavities and cracks and generating cavity crack analysis results; the underground risk detection module analyzes potential risks and environmental influences by adopting a risk assessment algorithm and an environmental influence analysis method based on the cavity crack analysis result, performs security risk assessment and generates an underground risk identification result; The geological risk dynamic evaluation module adopts a dynamic planning algorithm and a risk prediction model to evaluate the risk level and analyze the development trend based on the underground risk recognition result, and generates a geological risk dynamic evaluation result; the emergency path planning module generates an emergency obstacle avoidance path plan by adopting a path optimization algorithm and a safety navigation strategy based on a geological risk dynamic evaluation result; The detection route optimization module is used for adjusting detection strategies and optimization efficiency based on emergency obstacle avoidance path planning by adopting an efficiency analysis technology and a route re-planning method to generate an optimized detection route; the underground robot cooperation strategy module is used for carrying out task allocation and cooperation path planning by adopting a multi-robot coordination control algorithm and a communication synchronization technology based on the optimized detection route, so as to generate a robot cooperation operation plan.
2. The underground cavity disease detection robot control system of claim 1, wherein the cleaned and standardized underground data comprises electromagnetic properties, rock types and ground water levels of a stratum, the three-dimensional model of the stratum structure comprises layer sequence, thickness and lithology changes of the underground stratum, the cavity crack analysis results comprise directionality of cracks, volumes of the cavities and stress distribution of the stratum, the underground risk identification results comprise potential landslide areas, collapse risks and environmental vulnerability, the geological risk dynamic evaluation results comprise risk level changes, areas with estimated influences and emergency response time, the emergency obstacle avoidance path planning comprises coordinate of obstacle avoidance points, curvature of paths and time cost analysis, the optimized detection paths comprise improved detection depth, path coverage area and energy consumption estimation, and the robot cooperation operation plan comprises communication frequency, operation time sequence and distance control among robots.
3. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the underground data acquisition module comprises a sensor network sub-module, a data preprocessing sub-module and an environment monitoring sub-module; The sensor network submodule deploys a multipoint distributed sensor network based on an underground detection environment, performs data transmission by using Bluetooth or Wi-Fi, maintains the consistency and accuracy of data collected by a plurality of sensors in time by using a time synchronization protocol, and performs preliminary integration on collected stratum data and temperature and humidity information by weighted average or maximum likelihood estimation to generate an original underground data set; The data preprocessing submodule is used for classifying data by adopting a decision tree or a support vector machine based on an original underground data set, carrying out standardized processing on the data by using Min-Max Scaling, removing high-frequency noise in the data by using a low-pass filter, executing feature extraction operation, extracting key data features by using a principal component analysis method, and generating an optimized stratum data set; The environment monitoring submodule is used for carrying out deep analysis on environmental parameters by using regression analysis or variance analysis based on an optimized stratum data set, evaluating stratum stability and temperature and humidity change, predicting future environmental change trend by using time sequence analysis or a machine learning model, and generating cleaned and standardized underground data.
4. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the geological structure analysis module comprises a data modeling sub-module, a structure analysis sub-module and a three-dimensional visualization sub-module; The data modeling module adopts a geological modeling algorithm based on the cleaned underground data, uses a Scikit-learn library in Python to create K-means clusters, sets the number of clusters as 5, initializes a central point to select K-means++, sets the maximum iteration number as 300, classifies stratum data, uses a Pandas library to perform data preprocessing, comprises missing value processing and data type conversion, and generates a geological data preliminary model; The structure analysis submodule is based on a geological data preliminary model, adopts a structure analysis technology, uses a graph theory algorithm, creates a graph model in Python through NetworkX libraries, sets nodes as characteristic points in stratum data, edges are relations between stratum, weights of the edges are calculated based on similarity between stratum layers, uses a Bayesian network to perform stability assessment, uses pgmpy libraries to construct a network, sets a conditional probability table, searches an optimal network structure by using Hill Climbing algorithm, and generates a stratum structure analysis result through structure and parameter learning of the model; The three-dimensional visualization submodule performs three-dimensional visualization processing based on a stratum structure analysis result, uses three-dimensional modeling software Blender, imports data in the stratum structure analysis result, utilizes a Python script to automatically construct a stratum model in the Blender, sets materials and textures of stratum, adjusts transparency and color to distinguish different stratum, performs user interaction through scaling and rotation model operation, performs scene rendering by using a rendering engine built in the Blender, and generates a stratum structure three-dimensional model.
5. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the cavity and crack analysis module comprises a feature recognition sub-module, a cavity positioning sub-module and a crack analysis sub-module; The feature recognition submodule is used for carrying out feature extraction based on a stratum structure three-dimensional model, a pattern recognition algorithm is used for carrying out image feature recognition, a convolutional neural network is selected for carrying out image feature recognition, a CNN model is constructed to comprise three convolutional layers, each layer of activation function uses ReLU, the maximum pooling layer is set to be 2x2, an Adam optimizer and a cross entropy loss function are adopted, a fit method is used for training the model, a evaluate method is used for verifying the performance of the model, and a feature information diagram is generated; The cavity positioning sub-module is used for performing space positioning by using a geographic information system technology based on a characteristic information diagram by adopting a space analysis algorithm, comprises a space analysis tool of an ArcGIS or QGIS platform, comprises buffer area analysis and space superposition, sets the radius of a buffer area to be determined according to stratum characteristics, identifies the relative position of a cavity by space superposition operation, calculates the geometric characteristics of the cavity by utilizing a self-defined algorithm, comprises the central coordinates, the area and the volume of the cavity, establishes a three-dimensional model of the cavity by using Delaunay triangulation, and generates a cavity position result; The crack analysis submodule is used for positioning a crack by adopting a crack identification technology based on a characteristic information graph and a hole position result, adopting a Canny edge detection algorithm, setting a low threshold and a high threshold to be 50 and 150, extracting a crack profile, then adopting a crack width measurement method, comprising the steps of using an expansion and erosion morphological operation algorithm, setting a structural element size to be 3x3 pixels, measuring the crack width, quantifying the length and the direction of the crack by an automatic algorithm, evaluating the influence of the crack on the stability of the stratum by combining a stratum stability model, and generating a hole crack analysis result.
6. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the underground risk detection module comprises a risk assessment sub-module, an environmental impact analysis sub-module and a safety alarm sub-module; The risk assessment submodule carries out risk level prediction by adopting a logistic regression model based on the cavity crack analysis result, uses a Scikit-learn library of Python to execute logistic regression, sets the parameter regularization strength as 1.0 and a solver as liblinear when initializing the model, trains the model by using a fit method, carries out risk level prediction on new data by using a predict method, and generates a risk level prediction result; The environment influence analysis submodule performs environment influence analysis based on a risk level prediction result, performs Kriging spatial interpolation by adopting gstat packages of R language, sets interpolation parameters including model automatic fitting and a variation function type as a spherical model, generates a variation function model by utilizing a variogram function, performs spatial interpolation by applying a krige function, analyzes potential influence of crack risks on surrounding environment, and generates environment influence assessment; the safety alarm sub-module executes safety risk assessment by adopting a decision tree algorithm based on environmental impact assessment, creates a decision tree by using a Scikit-learn library of Python, sets the maximum depth of the tree to be 5 during initialization, sets the splitting standard as information gain, applies a fit method training model, analyzes environmental impact and risk level by using a predict method, formulates early warning measures and generates an underground risk identification result.
7. The system for controlling the robot for detecting the disease of the underground cavity according to claim 1, wherein the geological risk dynamic evaluation module comprises a risk prediction sub-module, a dynamic monitoring sub-module and a strategy generation sub-module; The risk prediction submodule carries out risk prediction by adopting a Bayesian network model based on an underground risk recognition result, defines nodes of differential risk types, including geological structures and historical risk events, sets the dependency relationship and conditional probability among the nodes, calculates the probability of occurrence of multiple types of risks through Bayesian reasoning, and generates a risk prediction analysis result; the dynamic monitoring sub-module is used for deploying a geological risk dynamic monitoring system based on a risk prediction analysis result, monitoring geological data in real time, collecting and analyzing key indexes including seismic wave speed and soil humidity in real time by using APACHE KAFKA, analyzing a risk development trend by using a time sequence algorithm, and generating real-time risk monitoring data; The strategy generation submodule carries out risk level assessment by applying a dynamic programming algorithm based on real-time risk monitoring data, defines state variables and decision variables in the process, including risk levels and countermeasures, sets objective functions to minimize risk influence, calculates optimal decisions under various situations, and generates a geological risk dynamic assessment result.
8. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the emergency path planning module comprises an obstacle avoidance path design sub-module, a safety navigation sub-module and a path optimization sub-module; The obstacle avoidance path design submodule carries out path design based on a geological risk dynamic evaluation result, adopts a geographic information system technology and an obstacle avoidance algorithm, carries out topography analysis and obstacle positioning by the geographic information system technology, uses a Dijkstra algorithm to carry out path search, sets parameters including a distance between nodes and a cost function, uses a Voronoi diagram to divide a safety area, assists path planning, and generates a preliminary obstacle avoidance path; The safety navigation sub-module utilizes real-time monitoring data based on a preliminary obstacle avoidance path, applies a safety navigation strategy, wherein data sources comprise sensor feedback and weather change, dynamically updates terrain information, uses an A-type algorithm to carry out path adjustment, sets parameters comprising current position, target point and obstacle information, matches environment change, and generates a real-time adjustment path; The path optimization submodule is used for generating an emergency obstacle avoidance path plan based on real-time path adjustment and applying a path optimization algorithm including a genetic algorithm and a simulated annealing algorithm, setting parameters including path length, time cost and energy consumption, and executing path optimization.
9. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the detection route optimizing module comprises a detection efficiency analyzing sub-module, a route re-planning sub-module and a detection strategy adjusting sub-module; The detection efficiency analysis submodule performs efficiency analysis based on detection task and environmental data by using a cluster analysis algorithm, the cluster analysis algorithm adopts a K-Means algorithm, the number of clusters is set to be 5, the iteration times is 100, the space coordinates and the time stamp are used as input parameters, euclidean distance is used as similarity measurement, the abnormal point detection uses a Z-Score method, the mean value and standard deviation of detection data are used as parameters, and data points which deviate from the mean level by more than 2 standard deviations are identified and used as abnormal points, so that an efficiency analysis result is generated; the route re-planning submodule performs route re-planning by using an A-search algorithm and a genetic algorithm based on an efficiency analysis result, wherein a heuristic function of the A-search algorithm is set to be Manhattan distance, the population size of the genetic algorithm is set to be 50 according to the obstacle and the route length, the cross probability is 0.7, the variation probability is 0.1, and 100 generations of iteration are performed to achieve optimal route configuration, and an adjusted route is generated; the detection strategy adjustment submodule adjusts detection parameters and strategies based on the adjusted route, the sensor sensitivity adjustment adopts a dynamic threshold adjustment algorithm, the threshold adjustment range is set to be +/-20% according to real-time environment changes, the detection frequency adjustment is carried out according to the complexity of the route and the expected detection duration, the frequency adjustment range is set to be 1-10 times per second, and the optimized detection route is generated.
10. The system for controlling a robot for detecting an underground cavity disease body according to claim 1, wherein the underground robot cooperation strategy module comprises a task allocation sub-module, a communication synchronization sub-module and a cooperation path planning sub-module; The task allocation submodule adopts a multi-robot coordination control algorithm based on an optimized detection route, and comprises initializing a robot state, including position coordinates, capability indexes and a task state, selects a robot based on fitness by using a roulette selection method, sets a cross probability parameter as 0.7, sets a variation probability parameter as 0.1, generates current task allocation, updates a task priority queue according to task urgency and the robot state, and generates a task allocation scheme; the communication synchronization submodule is based on a task allocation scheme, a communication synchronization technology is applied, a data packet format is set to comprise a sender ID, a message type and content, a data transmission speed parameter is defined to comprise transmission data quantity per second, a synchronization precision parameter comprises time deviation tolerance, a heartbeat detection mechanism is adopted to maintain stable communication, a heartbeat interval is set to be 1 second, a fault detection threshold is set to comprise retransmission data when no response is received for three times continuously, and thus a real-time communication protocol is generated; The collaborative path planning submodule is based on a real-time communication protocol, a path planning algorithm is utilized, dynamic planning and A search are combined, cost function parameters including path length and obstacle adjacency are set in the dynamic planning algorithm, heuristic functions including Manhattan distance are set in the A search algorithm, input parameters include current position coordinates, target position coordinates and environment obstacle positions of a robot, a differentiated path is estimated through the cost function, a smoothing coefficient is set to be 0.5 through path smoothing processing, obstacle avoidance distance parameters are set to be 0.5m during obstacle avoidance processing, and therefore a robot collaborative operation plan is generated.

Description

Underground cavity disease body detection robot control system Technical Field The invention relates to the technical field of underground detection, in particular to a robot control system for detecting an underground cavity disease body. Background The field of subsurface exploration technology has focused on developing and applying various methods and tools for subsurface exploration and analysis. The underground detection technology is widely applied to the fields of geological exploration, soil research, underground water source detection, mineral resource evaluation, environment monitoring and the like. In engineering construction, subsurface detection techniques are also of paramount importance, such as for detecting the position and status of underground pipes, cables or other facilities, as well as for assessing the stability and safety of underground structures. These techniques include seismic wave detection, electromagnetic detection, gravity and magnetic detection, geological radar detection, etc., all utilizing different physical principles to probe the structure and characteristics of the subsurface. The underground cavity disease body detection robot control system is an advanced technical system specially designed for identifying and analyzing underground cavities and disease bodies thereof (such as cracks, cavities and other abnormal structures affecting the safety of underground structures). The main purpose of this system is to improve the accuracy and efficiency of subsurface detection while reducing the risk and cost due to unknown conditions of the subsurface. By using robots driven by such control systems, detection can be made in situations where security of personnel poses a threat, or where manual access is difficult. Conventional systems exhibit several drawbacks in practical operation. The accuracy and efficiency of data acquisition are limited by the sensor network and the data processing method, so that the quality and usability of the data are affected, and the accuracy of analysis is further affected. The geological modeling and visualization technology is simple, complex stratum structures are difficult to effectively display, and accuracy of geological feature interpretation is limited. The lack of efficient crack identification and risk assessment techniques limits potential risk identification and environmental impact analysis, increasing the risk of underground operations. Under the emergency condition, the path planning and the route detection are insufficient in optimizing capability, and the emergency response efficiency and safety are affected. The lack of an effective multi-robot cooperation strategy limits the operation cooperation efficiency and influences the overall operation effect. Disclosure of Invention The invention aims to solve the defects in the prior art and provides a robot control system for detecting underground cavity disease bodies. In order to achieve the aim, the control system of the underground cavity disease body detection robot adopts the following technical scheme that the control system comprises an underground data acquisition module, a geological structure analysis module, a cavity and crack analysis module, an underground risk detection module, a geological risk dynamic evaluation module, an emergency path planning module, a detection route optimization module and an underground robot cooperation strategy module; The underground data acquisition module is based on an underground detection environment, adopts a sensor network technology and a machine learning preprocessing method, collects stratum data and temperature and humidity information, performs data filtering and feature extraction, performs normalization processing on the data, and generates cleaned and standardized underground data; the geological structure analysis module is used for creating a three-dimensional model of a subsurface stratum based on the cleaned and standardized underground data by adopting a geological modeling algorithm and a three-dimensional visualization technology, analyzing the distribution and composition characteristics of the stratum, and generating a stratum structure three-dimensional model; The cavity and crack analysis module is used for identifying underground cavities and cracks based on a stratum structure three-dimensional model by adopting a geological feature analysis method and a crack identification technology, evaluating the sizes and the shapes of the underground cavities and cracks and generating cavity crack analysis results; the underground risk detection module analyzes potential risks and environmental influences by adopting a risk assessment algorithm and an environmental influence analysis method based on the cavity crack analysis result, performs security risk assessment and generates an underground risk identification result; The geological risk dynamic evaluation module adopts a dynamic planning algorithm and a risk prediction