Autonomous navigation system and method for field inspection robot of buried pipeline

The autonomous navigation system for buried pipeline inspection robots addresses inefficiencies in existing methods by using electromagnetic holographic detection and sensor fusion to create visualization maps, enhancing detection efficiency and accuracy while minimizing pipeline damage.

CN114879699BActive Publication Date: 2024-12-06NORTHEASTERN UNIV CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210670764.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-15
Publication Date
2024-12-06
Estimated Expiration
2042-06-15

AI Technical Summary

Technical Problem

Existing pipeline inspection methods, whether internal or external, face limitations such as pipeline blockage, material restrictions, low detection efficiency, and human subjectivity, leading to inefficient and inaccurate inspections, especially in complex outdoor environments.

Method used

An autonomous navigation system for a buried pipeline inspection robot equipped with a host computer visualization data processing module, external environment multi-sensor perception module, multi-sensor information interaction storage module, bottom-level drive execution module, and electromagnetic holographic detection module, enabling ground-based inspection with electromagnetic holographic detection and sensor data fusion to create visualization maps and motion planning paths.

Benefits of technology

The system enhances detection efficiency, reduces pipeline damage, improves mobility, and minimizes human subjectivity, allowing for accurate and efficient inspection of buried pipelines with a detection range of up to 1.5 meters.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114879699B_ABST
    Figure CN114879699B_ABST
Patent Text Reader

Abstract

The present invention provides an autonomous navigation system and method for a field inspection robot for buried pipelines. The system comprises a host computer visualization data processing module, an external environment multi-sensing module, a multi-sensor information interaction storage module, a bottom-level driving execution module, and an electromagnetic holographic detection module. First, an above-ground and underground electronic map corresponding to an inspection area is established, and the obtained electronic map is initialized and set. A motion planning strategy for the inspection robot is established and solved according to the landmark node information in the electronic map, and an inspection robot model is established to view the robot posture information. After the inspection is completed, the inspection robot returns or goes to the next inspection area. The present invention proposes a method for establishing an above-ground and underground electronic map model for field buried pipelines, establishes a motion planning strategy for the field inspection robot based on the map model, and realizes a "patrol-inspect-storage-update" working mode of the inspection robot for buried pipelines in the field.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of inspection robots, and in particular relates to an autonomous navigation system and method for a field inspection robot for buried pipelines. Background Art

[0002] The existing pipeline inspection methods are generally divided into the following categories: 1) using an in-pipeline inspection robot to inspect the buried pipeline, and connecting a cable at the tail of the inspection robot to assist the in-pipeline inspection robot in performing the inspection task. 2) an out-pipeline inspection robot, which needs to be adsorbed on the surface of the pipeline, close to the pipe wall and move along the curve on the pipe wall to complete the pipeline inspection. At the same time, this type of robot has relatively strict requirements on the material of the pipeline wall. 3) manual inspection, the commonly used manual inspection methods are DC voltage gradient method (DCVG) and AC potential gradient method (ACVG). The inspection worker holds the pipeline inspection equipment and moves along the direction of the pipeline with the direction of the buried pipeline marked in advance. The inspection worker needs to stop and insert the equipment into the ground to inspect the buried pipeline every few to ten meters. At the same time, the existing inspection robots do not have the ability to perceive and inspect unknown areas in the field in complex outdoor environments. Therefore, the current field inspection technology of pipelines is mainly based on human inspection. The human inspection method has the disadvantages of strong subjectivity, low detection efficiency, and low detection accuracy.

[0003] The above detection methods each have their own characteristics, but their respective shortcomings are also extremely obvious. Method 1) Use an internal detection robot. The internal detection robot is inconvenient to move in the pipeline and is prone to pipeline blockage. In addition, the internal detection robot is affected by itself and the pipeline it is in, and its battery capacity and cable length are limited. At the same time, the pipeline needs to be stopped when the internal detection robot is inspecting, which will cause a certain degree of economic losses. Method 2) The external detection robot needs to adsorb the pipeline wall to inspect the pipeline. The inspection speed is low, and there are strict requirements on the material of the pipeline. Regarding method 3), the existing inspection robots do not have the ability to perceive and inspect in unknown areas in the wild, so the current field inspection technology of pipelines is still mainly based on manual inspection. This type of inspection method requires high experience from the operator. Due to the subjectivity of the inspection, there is no unified inspection standard, and the inspection speed is slow. Summary of the invention

[0004] Based on the above problems, the present invention provides an autonomous navigation system for a buried pipeline field inspection robot, including: a host computer visualization data processing module, an external environment multi-sensor perception module, a multi-sensor information interaction storage module, a bottom-level drive execution module, and an electromagnetic holographic detection module;

[0005] The external environment multi-sensing module is used to sense the ground surrounding environment information and generate surrounding environment image data and radar point cloud data;

[0006] The electromagnetic holographic detection module is used to collect electromagnetic information of the secondary magnetic field induced by the buried pipeline and generate electromagnetic data;

[0007] The bottom-level driving execution module is used to drive the crawler chassis of the inspection robot;

[0008] The multi-sensor information interaction storage module is used to receive and upload surrounding environment image data, radar point cloud data, and electromagnetic data;

[0009] The host computer visualization data processing module is used to construct a visualization map of the inspection area and a simulation model of the inspection robot, and generate a motion planning path for the inspection robot;

[0010] The external environment multi-sensing module includes a monocular industrial camera and a 3D laser radar; the multi-sensor information interaction storage module is a main control board chip, the monocular industrial camera is fixed in the middle position of the front side of the crawler chassis cover of the inspection robot, and the 3D laser radar is fixed in the middle position of the upper side of the crawler chassis cover of the inspection robot through a radar rack; the monocular industrial camera and the 3D laser radar are electrically connected to the input end of the main control board chip respectively; the image data of the surrounding environment on the ground is collected by the monocular industrial camera, and the point cloud data of the surrounding obstacles is collected by the 3D laser radar;

[0011] The electromagnetic holographic detection module includes an electromagnetic holographic detector and a mechanical arm; the mechanical arm is fixed to the upper tail of the crawler chassis cover of the inspection robot, the mechanical arm clamps a probe with an excitation coil, and is used to transmit electromagnetic signals to the underground pipeline; the electromagnetic holographic detector is installed at the lower front side of the cover of the crawler chassis of the inspection robot, and receives electromagnetic data at a certain angle to the horizontal cover, and collects electromagnetic data of the secondary magnetic field induced by the underground pipeline through the electromagnetic holographic detector;

[0012] The bottom-level drive execution module is a drive control board installed on the crawler chassis of the inspection robot. The control end of the drive control board is connected to the servo motor of the inspection robot, and the controlled end of the drive control board is connected to the main control board chip via CAN_TO_USB communication.

[0013] The autonomous navigation system also includes a power supply module, which includes a ternary lithium battery pack and a battery voltage conversion module installed on the track chassis. The battery voltage conversion module converts the voltage output by the ternary lithium battery pack into the driving voltage required by each module.

[0014] An autonomous navigation method for a buried pipeline field inspection robot is implemented based on the autonomous navigation system for a buried pipeline field inspection robot, and the autonomous navigation method includes:

[0015] Step 1: Start the field buried pipeline inspection robot and control the inspection robot to move to the base point of the area to be inspected;

[0016] Step 2: Create a visual electronic map of the corresponding inspection area and initialize the electronic map;

[0017] Step 3: Based on the map information with the underground pipeline direction currently obtained by the inspection robot, establish and solve the motion planning strategy of the inspection robot;

[0018] Step 4: In the visual data processing module of the host computer, establish a simulation model of the inspection robot and view the relevant status operation data of the robot;

[0019] Step 5: After the inspection, the inspection robot returns or goes to the next inspection area;

[0020] The step 2 comprises:

[0021] Step 2.1: Set the inspection area range parameters of the field inspection robot; the default setting of the inspection area is a rectangle. Setting the inspection area requires two parameters, namely the side length L of the inspection area and area and the side width W of the inspection area area , the value range of both is 0<L area , W area ≤400;

[0022] Step 2.2: Start the inspection robot to build ground environment map and underground pipeline environment map models;

[0023] The process of building the ground environment map model is as follows:

[0024] Step A1: The host computer establishes a communication connection with the inspection robot through remote SSH (Secure Shell, SSH for short). After the connection is established, a control command is issued, and the inspection robot moves to the inspection area according to the control command; the monocular industrial camera ROS node is started, and the startup command is: roslaunchrobot_mappingcamara_mapping.launch; the camera sampling frame rate is set in the monocular industrial camera startup node, and the first image captured by the camera is used as the base point P of the inspection area coordinate system world =(x world ,y world ,z world );

[0025] Step A2: The monocular industrial camera collects the feature point differences between different images at adjacent moments and calculates the camera's position P in the inspection area coordinate system through triangulation. camera =(x camera ,y camera ,z camera );

[0026] Step A3: According to the static coordinate relationship between the camera and the chassis of the inspection robot, issue the static coordinate transformation command: roslaunch tf2_ros static_transform_publisher / robot / camera to obtain the position coordinates (x robot ,y robot ,z robot ,θ robot ), θ robot is the yaw angle of the inspection robot in the world coordinate system;

[0027] Then the posture equation of the inspection robot is expressed as: k =f(x k-1 ,u k ,w k )=P 1 x k-1 +Q 1 u k +w k , P 1 is the posture transformation matrix of the inspection robot between adjacent moments, Q 1 is the coordinate transformation coefficient, w k It is the noise interference during the movement of the inspection robot.

[0028] The observation equation of the inspection robot is expressed as: k,j =h(y j ,x k ,v k,j )=P 2 y j +Q 2 x k +v k,j , P 2 is the measurement matrix of the feature points, Q 2 is the coefficient of the equation of motion, v k,j The error caused by observing the ground environment.

[0029] Step A4: Start the inspection robot 3D lidar ROS node, the startup command is: roslaunchrobot_mapping lidar_mapping.launch. After the radar node is started, it receives the three-dimensional point cloud data P of the inspection robot's surrounding environment and obstacles relative to the radar coordinate system. lidar =(x lidar ,y lidar ,z lidar );

[0030] Step A5: According to the static coordinate relationship between the radar and the chassis of the inspection robot, issue the static coordinate conversion command: roslaunch tf2_ros static_transform_publisher / robot / lidar to obtain the coordinates P of the surrounding targets in the inspection area coordinate system. object =(x object ,y object ,z object ), and optimize the inspection robot posture data calculated by the camera node.

[0031] Get the ground map model of the inspection robot.

[0032] The process of building an underground pipeline environment map model is as follows:

[0033] Step B1: The inspection robot arm holds the probe with the excitation coil and moves it in the inspection area at V robot =Moving at a speed of 1m / s;

[0034] Step B2: Start the electromagnetic holographic detector ROS node, the startup command is: roslaunch robot_mappingelectromagnetic_mapping.launch;

[0035] Step B3: The ternary lithium battery pack applies an alternating current to the excitation coil through the power module, and sets the excitation coil voltage amplitude V in the holographic detector node. ExcitingCoil =24v, excitation coil current amplitude A ExcitingCoil =3A, frequency F ExcitingCoil =32Hz;

[0036] Step B4: After applying alternating current to the excitation coil, a primary magnetic field is induced on the coil, and the primary magnetic field enters the ground to induce a secondary magnetic field on the surface of the pipeline;

[0037] Step B5: During the robot's mobile inspection, the electromagnetic holographic detector receives the secondary magnetic field data. When the voltage minimum is detected, the current underground landmark observation point q is recorded. i Combined with the robot's current position, the electronic map model of the underground pipeline is obtained:

[0038] Buried pipeline observation equation: p k,i =g(q i ,x k ,e k,i )=(P 3 +d lift-off +r robot-arm )q i +Q 3 x k +e k,i , P 3 is the observation matrix of pipeline landmark nodes, d lift-off is the lift-off value, i.e., the vertical distance between the buried pipeline and the inspection robot, d lift-off ≤1.5, indicating that the maximum depth of buried pipelines that the inspection robot can detect is 1.5 meters, r robot-arm is the detectable area of ​​the inspection robot centered on the robotic arm, r robot-arm ≤1.304, indicating that the maximum detection range of the inspection robot is 1.304 meters, Q 3 is the coefficient of the current robot motion equation, e k,i To detect the errors in the underground pipeline environment;

[0039] Start the sensor map summary matching ROS node: roslaunchrobot_mapping sensor_mapping.launch, in which the coordinate information and timestamp information of the above-ground and underground nodes are used to match the above-ground and underground landmark nodes;

[0040] Start the sensor state publishing node: roslaunchjoint_state_publisherjoint_state_publisher.launch;

[0041] Start the robot state publishing node: roslaunchrobot_state_publisherrobot_state_pblisher.launch;

[0042] Start Rviz on the host computer, create a map component in Rviz, set the subscribed topic name in the component to / map, and view the visualized ground environment electronic map generated by the two nodes.

[0043] When the obtained electronic map contains a ground environment map model and an underground pipeline environment map model, execute steps 2.2 to 2.5 to finally obtain a landmark node sequence in which the landmark nodes of the underground pipeline are sorted from small to large;

[0044] When the obtained visualization map only has the ground environment map model but no underground pipeline environment map model, the inspection area range parameters set in step 1 are used as the standard. In this inspection area, the inspection robot takes the set inspection area base point as the starting point and makes an S-shaped curve movement along the length of the inspection area to detect underground pipelines. The radius of the curve for the inspection robot to turn is 0.25 meters. When an underground pipeline is found, a landmark node sequence is generated based on the current underground pipeline landmark node in the order of small to large timestamps.

[0045] When the obtained visualization map contains only the ground environment map model and no underground pipeline environment map model, and no underground pipeline is found while performing S-shaped curve movement along the length of the inspection area, it is considered that there is no underground pipeline in this inspection area, and the inspection robot stops moving and sends information to the upper computer to report that there is no underground pipeline here;

[0046] Step 2.3: Generate a landmark node list according to the landmark nodes of the above-ground environment map and the underground pipeline environment map; the landmark node list includes: above-ground and underground landmark node numbers, timestamps corresponding to each landmark node, and distances between landmark nodes;

[0047] Step 2.4: Using the timestamp of the landmark node generated by the inspection robot as the registration unit, match the above-ground and underground landmark nodes to obtain a set of landmark node sequences based on the underground pipeline landmark nodes;

[0048] Step 2.5: Arrange the obtained landmark node sequence in ascending order of timestamps and store it in the multi-sensor information interaction storage module;

[0049] The step 3 comprises:

[0050] Step 3.1: The inspection robot starts inspection from the node with the smallest landmark node number. During the inspection, the pipeline node number is processed according to the following two situations;

[0051] ① The inspection robot covers the buried pipeline during the inspection:

[0052] If the pipeline currently being inspected covers other surrounding pipelines, the inspection robot continues to inspect the pipeline normally according to the number in the landmark node sequence of the current pipeline;

[0053] If the current inspection pipeline is covered by other surrounding pipelines, a pipeline road sign node is added at the starting position of the current inspection pipeline, and a pipeline node number is also added at the end position of the current inspection pipeline. At the same time, the distance between the two covered nodes and the corresponding timestamp are calculated and saved. The current pipeline cannot be inspected within this coverage distance;

[0054] ② When the inspection robot finds a branch in the pipeline during the inspection of the current buried pipeline, the inspection robot sets the pipeline mark bit pipeFlagNo_x_y for each branch pipeline, and then sends a request signal to the host computer to request a response from the host computer. After receiving the request signal, the host computer tells the robot which branch pipeline to inspect based on the pipeline mark bit. The inspection robot regenerates the landmark node for the branch pipeline and performs the inspection normally.

[0055] Step 3.2: Search for all existing key path points between the current landmark node and the next landmark node and calculate the distance from the current landmark node to the next landmark node through the key path points;

[0056] Step 3.3: Sort the obtained distances, calculate the shortest distance and its corresponding key path point, and update the distance information of the two landmark nodes in the landmark node list;

[0057] Step 3.4: Control the inspection robot to move to the next landmark node at the shortest distance, and repeat step 3.3;

[0058] Step 3.5: Update the distance between each landmark node in the corresponding landmark node list and record it.

[0059] The beneficial effects of the present invention are:

[0060] 1. Compared with the in-pipeline inspection robot, the field inspection robot designed in the present invention runs on the ground and inspects the buried pipeline, which avoids the situation where the robot blocks the pipeline, reduces the damage to the pipeline caused by the inspection robot during operation, and can ensure the normal operation of the pipeline during the robot inspection process.

[0061] 2. Compared with the attached pipeline robot, the inspection robot designed by the present invention realizes 1.5m long-distance detection of pipelines through modular electromagnetic holographic technology and has better mobility.

[0062] 3. Compared with the current mainstream manual detection technology, the robot designed by the present invention has higher detection efficiency, reduces the influence of human subjectivity on buried pipeline detection and improves the detection speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a block diagram of the autonomous navigation system of the buried pipeline field inspection robot in the present invention;

[0064] Figure 2 It is a flow chart for initializing and setting the above-ground and underground electronic maps in the present invention;

[0065] Figure 3This is a flow chart for initializing the system described in the present invention;

[0066] Figure 4 This is a flow chart of the inspection robot detection in the present invention;

[0067] Figure 5 This is a flowchart of autonomous inspection of underground pipelines by the inspection robot in the present invention;

[0068] Figure 6 This is a flow chart of the inspection robot in the present invention determining the state of a buried pipeline;

[0069] Figure 7 The present invention is a flowchart of the autonomous navigation method for the buried pipeline field inspection robot. DETAILED DESCRIPTION

[0070] The invention will be further described below with reference to the accompanying drawings and specific implementation examples.

[0071] like Figure 1 As shown, an autonomous navigation system for a buried pipeline field inspection robot includes: a host computer visualization data processing module, an external environment multi-sensing module, a multi-sensor information interaction storage module, a bottom-level drive execution module, and an electromagnetic holographic detection module;

[0072] The external environment multi-sensing module is used to sense the ground surrounding environment information and generate surrounding environment image data and radar point cloud data;

[0073] The electromagnetic holographic detection module is used to collect electromagnetic information of the secondary magnetic field induced by the buried pipeline and generate electromagnetic data;

[0074] The bottom-level driving execution module is used to drive the crawler chassis of the inspection robot;

[0075] The multi-sensor information interaction storage module is used to receive and upload surrounding environment image data, radar point cloud data, and electromagnetic data;

[0076] The host computer visualization data processing module is used to construct a visualization map of the inspection area and a simulation model of the inspection robot, and generate a motion planning path for the inspection robot.

[0077] The external environment multi-sensing module includes a monocular industrial camera and a 3D laser radar; the multi-sensor information interaction storage module is a main control board chip, the monocular industrial camera is fixed in the middle position of the front side of the crawler chassis cover of the inspection robot, and the 3D laser radar is fixed in the middle position of the upper side of the crawler chassis cover of the inspection robot through a radar rack; the monocular industrial camera and the 3D laser radar are electrically connected to the input end of the main control board chip respectively; the image data of the surrounding environment on the ground is collected by the monocular industrial camera, and the point cloud data of the surrounding obstacles is collected by the 3D laser radar;

[0078] The multi-sensor information interaction storage module is a Raspberry Pi4b micro-control board chip based on ARM. It plays a connecting role and is mainly responsible for the following tasks: 1) Establishing data communication connection with the host computer through the SSH communication mechanism and feeding back data, 2) Performing data interaction and data fusion in each unit of the external environment multi-sensing module, 3) Connecting with the bottom-level drive control board of the inspection robot to realize feedback control of the bottom-level drive. It includes a data interaction processing unit, a map storage unit, a patrol robot motion planning path storage unit, and an inspection result storage unit; the data interaction processing unit is used to calculate the data between each sensor and the host computer, the map storage unit is used to store the generated electronic map, and the motion planning storage unit is used to save the path information generated after the search. In the next inspection, the map is checked first. If there is a map that has been inspected before, the electronic map and path rule can be directly called. The inspection result storage unit is used to save the pipeline status data obtained after the inspection of the buried pipeline.

[0079] The connection between the main control board chip and other devices is physical connection.

[0080] The external environment multi-sensing module includes a monocular industrial camera and a 3D laser radar. The monocular camera and the 3D laser radar are directly connected to the main control board chip through the USB 3.0 serial bus. In order to realize the real-time processing of the perceived data, the USB 3.0 port is used to directly connect the main control board chip and the sensor components to speed up the real-time data transmission and processing speed. The monocular industrial camera extracts features from the ground environment image, finds the landmarks in the ground environment, and then cooperates with the visual odometer to locate the inspection robot in the ground environment and build a map of the ground environment. At the same time, since a single camera is sensitive to the influence of light and weather, a laser radar is added to the system to assist the industrial camera in mapping the ground environment. At the same time, the laser radar can quickly generate point cloud data to calculate the distance between the robot and the surrounding objects, thereby saving calculation time.

[0081] The electromagnetic holographic detection module includes an electromagnetic holographic detector and a mechanical arm; the mechanical arm is fixed to the upper tail of the crawler chassis cover of the inspection robot, the mechanical arm clamps a probe with an excitation coil, and is used to transmit electromagnetic signals to the underground pipeline; the electromagnetic holographic detector is installed at the lower front side of the cover of the crawler chassis of the inspection robot, and receives electromagnetic data at a certain angle to the horizontal cover, and collects electromagnetic data of the secondary magnetic field induced by the underground pipeline through the electromagnetic holographic detector;

[0082] The electromagnetic holographic module includes an electromagnetic holographic detector and a mechanical arm. The mechanical arm is directly connected to the main control board chip, which is controlled by the main control board chip to assist the electromagnetic holographic detection board in detecting buried pipelines. The electromagnetic holographic detection board is connected to the drive control board through the crawler chassis of the inspection robot, and emits electromagnetic holographic detection signals to the ground at a certain angle. At the same time, the underground pipeline information obtained is processed. The signal form is as follows:

[0083] F=U c (·)+D rp (·)+V r (·)+σ(·)+M(·)+P(·)

[0084] Among them, F is the underground pipeline environment mechanism model based on electromagnetic holography, U c is the excitation coil voltage. The system excitation coil voltage range is 45V≤U c ≤350V, D rp In order to detect the distance, this system uses electromagnetic holography to detect underground pipelines, and its detection radius is D rp =10, unit is meter, V r is the speed of the inspection robot in the environment, σ is the stress on the buried pipeline from the surrounding environment, M is the electromagnetic interference generated by the surrounding environment, and P is the pipeline's own characteristics, including the inner diameter of the pipeline P r and pipe wall thickness P t .

[0085] The bottom-level drive execution module is a drive control board installed on the crawler chassis of the inspection robot. The control end of the drive control board is connected to the servo motor of the inspection robot, and the controlled end of the drive control board is connected to the main control board chip via CAN_TO_USB communication.

[0086] The bottom-level drive execution module is responsible for adjusting the speed and steering of the inspection robot and other drive control functions, and is connected to the main control board chip. The rotary encoder is connected to the drive control board through a DC reduction motor, and can measure the motor speed while cooperating with PWM technology to achieve motor speed regulation. The rotary encoder converts the motor angular velocity into a pulse number output, and the formula is as follows:

[0087] N output =g r *N encoder

[0088] Where N output is the number of pulses per rotation of the DC motor output shaft, g r is the motor gear ratio. In this system, g r =1:70, N encoder The number of pulses generated by the encoder during one rotation.

[0089] The speed formula is as follows:

[0090] m output =m input / g r

[0091] Among them, m output is the motor output speed, m input is the motor input speed, g r is the motor reduction ratio.

[0092] The autonomous navigation system also includes a power supply module, which includes a ternary lithium battery pack and a battery voltage conversion module installed on the track chassis. The battery voltage conversion module converts the voltage output by the ternary lithium battery pack into the driving voltage required by each module.

[0093] The upper computer visualization data processing module includes an industrial computer and a buried pipeline detection platform. The industrial computer is installed with Windows and Ubuntu dual systems, and the Ubuntu system forms a software connection with the Ros robot operating system, Gazebo, and Rviz. The inspection robot mapping algorithm and motion planning algorithm are developed in the Ros robot operating system. The inspection robot simulation model can be established in the Ros robot operating system by writing Xacro files. The Xacro file format is xml. Then the constructed Xacro simulation model is imported into Gazebo, and the environment or temporary obstacle information is added to Gazebo to perform dynamic simulation of the inspection robot, so that the motion state of the inspection robot under the current environmental information can be viewed. The constructed Xacro inspection robot simulation model is imported into Rviz, and the corresponding sensor component model is added to Rviz, so that the data collected by the sensor and the posture information of the inspection robot can be viewed. Finally, the electronic map model file is imported, and the Xacro inspection robot simulation model, Gazebo, and Rviz are integrated into the launch file to realize the dynamic simulation and posture data monitoring of the inspection robot.

[0094] The host computer communicates with the multi-sensor information interaction storage module, receives the ground and underground environment perception data input collected by the module, and respectively establishes the ground environment electronic map and underground pipeline environment electronic map visualization model through the ground and underground environment perception data, imports the inspection robot model established above into the ground and underground electronic maps, and then obtains the motion state and posture data of the inspection robot in the current map space. The buried pipeline detection platform is the back-end processing unit of the system, which is responsible for receiving the input of the underground pipeline electronic map model and analyzing the status of the buried pipeline.

[0095] like Figure 7As shown, an autonomous navigation method for a buried pipeline field inspection robot is implemented based on the autonomous navigation system for a buried pipeline field inspection robot, and the method includes:

[0096] Step 1: Start the underground pipeline inspection robot in the field and control it to move to the base point of the inspection area. First, use the charger to charge the power module of the inspection robot. After charging is completed, the power status of the inspection robot is marked as sending a charging completion message to the upper computer. Then, the upper computer controls the inspection robot to go to the base point of the inspection area and establishes a three-dimensional world coordinate system (x world ,y world , z world ).

[0097] Step 2: Create a visual electronic map of the corresponding inspection area and initialize the electronic map; Figure 2 As shown, including:

[0098] Step 2.1: Set the inspection range of the field inspection robot. The maximum inspection area is set to a rectangle. The inspection area includes two parameters, namely the length L of the inspection area and the width W of the inspection area. At the same time, the product of the two parameters must be less than the maximum inspection area S of the robot. max 90%, that is:

[0099] L*W≤0.9S max

[0100] Step 2.2: Start the inspection robot to build the ground environment map and underground pipeline environment map model; including:

[0101] Start the sensors on the inspection robot, set the start flag of each sensor, and judge whether the sensors on the inspection robot are successfully started by judging the value of each sensor start flag. If the start is successful, report the success of the start to the host computer and upload the start flag of the corresponding sensor. If the start fails, continue to start 3 times in a row. If the start still fails, the inspection robot will upload the flag of the sensor that failed to start to the host computer. Figure 3 shown.

[0102] Start the crawler chassis of the inspection robot in the same way and determine whether the crawler chassis is started successfully. At the same time, upload the chassis startup flag to the host computer to report the startup status of the current crawler chassis of the inspection robot.

[0103] Establish electronic map models of the above-ground environment and the underground pipeline environment. The current electronic map model contains both the above-ground environment map model and the underground pipeline environment map model. The model construction mechanism is as follows:

[0104]

[0105] The electronic map designed above contains three equations in total, f(x k-1 ,u k , w k ) is the motion equation of the field inspection robot based on the ground environment, h(y j ,x k ,v k,j ) is the ground observation equation obtained by the field inspection robot based on the landmark information in the ground environment, g(q i ,x k ,e k,i ) is the underground environment observation equation obtained by the field inspection robot based on the underground pipeline environment data measured by the electromagnetic holographic module. The ground observation equation is matched and combined with the underground observation equation using the current inspection robot's posture data and underground landmark nodes while taking the timestamp as the matching reference.

[0106] Set the world coordinate system, robot base coordinate system, and coordinate system of each sensor element, and establish the world three-dimensional coordinate system (x world ,y world , z world ), the world coordinate system is named world, and the robot's three-dimensional base coordinate system (x robot ,y robot , z robot ), the inspection robot base coordinate system is named base_link, and the inspection robot coordinate system also includes the three-dimensional coordinate system of each external sensor. When the inspection robot inspects the buried pipeline in the field, the position change of the inspection robot and each sensor on the robot relative to the world coordinate system is obtained according to the coordinate transformation between the coordinate systems.

[0107] The process of building the ground environment map model is as follows:

[0108] Step A1: The host computer establishes a communication connection with the inspection robot through remote SSH (Secure Shell, SSH for short). After the connection is established, a control command is issued, and the inspection robot moves to the inspection area according to the control command; the monocular industrial camera ROS node is started, and the startup command is: roslaunchrobot_mappingcamara_mapping.launch; the camera sampling frame rate is set in the monocular industrial camera startup node, and the first image captured by the camera is used as the base point P of the inspection area coordinate system world =(x world ,y world ,z world );

[0109] Step A2: The monocular industrial camera collects the feature point differences between different images at adjacent moments and calculates the camera's position P in the inspection area coordinate system through triangulation. camera =(x camera ,y camera ,z camera );

[0110] Step A3: According to the static coordinate relationship between the camera and the chassis of the inspection robot, issue the static coordinate transformation command: roslaunch tf2_ros static_transform_publisher / robot / camera to obtain the position coordinates (x robot ,y robot ,z robot ,θ robot ), θ robot is the yaw angle of the inspection robot in the world coordinate system;

[0111] Then the posture equation of the inspection robot is expressed as: k =f(x k-1 ,u k ,w k )=P 1 x k-1 +Q 1 u k +w k , P 1 is the posture transformation matrix of the inspection robot between adjacent moments, Q 1 is the coordinate transformation coefficient, w k It is the noise interference during the movement of the inspection robot.

[0112] The observation equation of the inspection robot is expressed as: k,j =h(y j ,x k ,v k,j )=P 2 y j +Q 2 x k +v k,j , P 2 is the measurement matrix of the feature points, Q 2 is the coefficient of the equation of motion, v k,j The error caused by observing the ground environment.

[0113] Step A4: Start the inspection robot 3D lidar ROS node, the startup command is: roslaunchrobot_mapping lidar_mapping.launch. After the radar node is started, it receives the 3D point cloud data P of the inspection robot's surrounding environment and obstacles relative to the radar coordinate system. lidar =(x lidar ,y lidar ,z lidar ).

[0114] Step A5: According to the static coordinate relationship between the radar and the chassis of the inspection robot, issue the static coordinate conversion command: roslaunch tf2_ros static_transform_publisher / robot / lidar to obtain the coordinates P of the surrounding targets in the inspection area coordinate system. object =(x object ,y object ,z object ), and optimize the inspection robot posture data calculated by the camera node.

[0115] Get the ground map model of the inspection robot.

[0116] The process of building an underground pipeline environment map model is as follows:

[0117] Step B1: The inspection robot arm holds the probe with the excitation coil and moves it in the inspection area at V robot =Moving at a speed of 1m / s;

[0118] Step B2: Start the electromagnetic holographic detector ROS node, the startup command is: roslaunch robot_mappingelectromagnetic_mapping.launch;

[0119] Step B3: The ternary lithium battery pack applies an alternating current to the excitation coil through the power module, and sets the excitation coil voltage amplitude V in the holographic detector node. ExcitingCoil =24v, excitation coil current amplitude A ExcitingCoil =3A, frequency F ExcitingCoil =32Hz;

[0120] Step B4: After applying alternating current to the excitation coil, a primary magnetic field is induced on the coil, and the primary magnetic field enters the ground to induce a secondary magnetic field on the surface of the pipeline;

[0121] Step B5: During the robot's mobile inspection, the electromagnetic holographic detector receives the secondary magnetic field data. When the voltage minimum is detected, the current underground landmark observation point q is recorded. i Combined with the robot's current position, the electronic map model of the underground pipeline is obtained:

[0122] Buried pipeline observation equation: p k,i =g(q i ,x k ,e k,i )=(P 3 +d lift-off +r robot-arm )q i +Q 3 x k +e k,i , P 3 is the observation matrix of pipeline landmark nodes, d lift-off is the lift-off value, i.e., the vertical distance between the buried pipeline and the inspection robot, d lift-off ≤1.5, indicating that the maximum depth of buried pipelines that the inspection robot can detect is 1.5 meters, r robot-arm is the detectable area of ​​the inspection robot centered on the robotic arm, r robot-arm ≤1.304, indicating that the maximum detection range of the inspection robot is 1.304 meters, Q 3 is the coefficient of the current robot motion equation, e k,i To detect the errors in the underground pipeline environment;

[0123] Start the sensor map summary matching ROS node: roslaunchrobot_mapping sensor_mapping.launch, in which the coordinate information and timestamp information of the above-ground and underground nodes are used to match the above-ground and underground landmark nodes;

[0124] Start the sensor state publishing node: roslaunchjoint_state_publisherjoint_state_publisher.launch;

[0125] Start the robot state publishing node: roslaunchrobot_state_publisherrobot_state_pblisher.launch;

[0126] Start Rviz on the host computer, create a map component in Rviz, set the subscribed topic name in the component to / map, and view the visualized ground environment electronic map generated by the two nodes.

[0127] The first two terms of the above equations in the traditional ground electronic map model, where k = 1, 2, ..., K is the time corresponding to the operation of the inspection robot, x k represents the position and posture of the inspection robot at time k, that is, the three-dimensional coordinates (x robot ,yrobot , z robot ,θ robot ), θ robot The inspection robot deviates from the world coordinate system by an angle of 0°<θ robot ≤360°,u k is the motion control input, i.e., the change in the inspection robot’s posture u from time k to time k+1 k =(Δx robot , Δy robot , Δz robot , Δθ robot ), y j is the landmark node of the ground environment, that is, N feature information y extracted from the ground environment j =(y 1 ,y 2 , ..., y N ), z k,j Indicates that the inspection robot is at x k The ground landmark node y is observed under the pose j The observation data generated when z k,j =(z 1 k,j , z 2 k,j , ..., z m k,j ).

[0128] The buried pipeline information is added to the traditional ground environment map model to form an underground pipeline environment map model, where q i is the underground pipeline environment landmark observation node, q i =(q 1 ,q 2 , ..., q N ),p k,i Indicates the inspection robot in its own position x k The underground pipeline environment landmark node q is observed i The detection data generated when p k,j =(p 1 k,j , p 2 k,j , ..., p m k,j ), w k is the noise interference during the movement of the inspection robot, v k,j is the error generated in the process of observing the ground environment, e k,i To detect the errors in the underground pipeline environment.

[0129] It is difficult to directly solve this conditional distribution, so factor graph optimization is used for state estimation:

[0130]

[0131] Among them, r represents the residual between the observation model and the actual observation, ∑ 0 ,∑ pk ,∑ p(m,n) Respectively represent the corresponding covariance matrix, r 0 Corresponding to the initial value x of x 0 The remaining items will be described in detail below.

[0132] In the process of factor graph optimization, it is necessary to insert the map registration factor, corresponding to

[0133]

[0134] in, represents the relative posture relationship, and Θ represents the inverse operation of the posture transformation in se(3).

[0135] Then insert the loop factor into the factor graph, corresponding to

[0136]

[0137] Whenever a new node is inserted into the factor graph, the entire factor graph is optimized and calculated. The factor graph is an incremental processing backend optimization, which avoids redundant calculations. After the robot's own position x is obtained, the underground environment map can be constructed through the conversion relationship between the underground environment and the robot's position:

[0138] q i =x k +TRp k,i

[0139] Among them, R is the mapping deflection matrix, which is related to the robot's current posture. When the robot is in a horizontal state, R is the unit matrix, and T is the homogeneous mapping transformation matrix between the underground observation p and the actual robot position x.

[0140]

[0141] Wherein, s is the map zoom factor, and s is an integer in this system.

[0142] In general, the above-ground and underground environments are jointly matched based on timestamps through the robot's position x, and finally a map model of the above-ground and underground environments is formed.

[0143] In the process of establishing the above-ground and underground electronic maps, the following situations may occur: Figure 4 To further explain:

[0144] ① When the obtained electronic map contains the above-ground environment model and the underground pipeline environment model, execute steps 2.2 to 2.5, and finally obtain a sequence of landmark nodes in which the landmark nodes of the underground pipeline are sorted from small to large;

[0145] ② When the electronic map obtained only contains the ground environment model but no underground pipeline environment model, the inspection area range parameters set in step 1 are used as the standard, and the inspection robot performs S-shaped curve movement along the length of the inspection area to detect underground pipelines. The radius of the curve for the inspection robot to turn is 0.25 meters. When an underground pipeline is found, the landmark node sequence is generated based on the current underground pipeline landmark node in the order of the timestamp from small to large.

[0146] ③ When the obtained electronic map only contains the ground environment model but no underground pipeline environment, and the underground pipeline environment model cannot be obtained after step ②, there is no underground pipeline in the inspection area. The inspection robot stops moving and sends information to the upper computer to report that there is no underground pipeline here;

[0147] Step 2.3: Generate a landmark node sequence according to the landmark nodes of the ground map and the underground pipeline environment map, wherein the landmark node sequence includes the ground landmark node number, the underground landmark node number, the timestamp corresponding to the node, and the distance between the landmark nodes;

[0148] Step 2.4: For the underground pipeline marking position, the timestamp of the landmark node generated by the inspection robot is used as the registration unit, and the above-ground and underground landmark nodes are matched to obtain a set of landmark node sequences based on the underground pipeline landmark nodes;

[0149] Step 2.5: Sort the obtained landmark node sequence in ascending order of timestamps and store it in the multi-sensor information interaction storage module.

[0150] Step 3: Based on the map information with the underground pipeline direction currently obtained by the inspection robot, establish the motion planning strategy of the inspection robot and solve it, such as Figure 3 As shown;

[0151] 1) Determine the direction of the underground pipeline according to the underground pipeline landmark node sequence. Since the numbering of the underground landmark nodes is generated along the pipeline, the pipeline numbers in the underground pipeline landmark node sequence from small to large represent the direction of the underground pipeline.

[0152] 2) Since there may be more than one underground pipeline in the constructed electronic map, it is necessary to set a pipeline mark position, pipeFlagNo_x, for each underground pipeline when constructing the underground pipeline map, and the host computer tells the inspection robot which pipeline to inspect.

[0153] 3) The inspection robot finds the node with the smallest number in the underground pipeline node list corresponding to the pipeline mark position according to the pipeline mark position set by the host computer as the starting position of the robot's inspection of the current pipeline, and starts the inspection; Figure 5 As shown in the figure, the inspection process is as follows:

[0154] Step 3.1: The inspection robot starts inspection from the node with the smallest landmark node number;

[0155] Step 3.2: Search for all existing key path points between the current landmark node and the next landmark node and calculate the distance from the current landmark node to the next landmark node through the key path points;

[0156] Step 3.3: Sort the obtained distances, calculate the shortest distance and its corresponding key path point, and update the distance information of the two landmark nodes in the landmark node list;

[0157] Step 3.4: Control the inspection robot to move to the next landmark node at the shortest distance, and repeat step 3.3;

[0158] Step 3.5: Update the distance between each landmark node in the corresponding landmark node list and record it.

[0159] The following situations may occur when the inspection robot inspects buried pipelines:

[0160] ① The inspection robot covers the buried pipeline during the inspection:

[0161] If the pipeline currently being inspected covers other surrounding pipelines, the inspection robot continues to inspect the pipeline normally according to the number in the landmark node sequence of the current pipeline.

[0162] If the current inspected pipeline is covered by other surrounding pipelines, a pipeline road sign node is added at the starting position of the currently inspected pipeline, and a pipeline node number is also added at the end position of the currently inspected pipeline. At the same time, the coverage distance between the two nodes is calculated and saved. The current pipeline cannot be inspected within this coverage distance.

[0163] ② When the inspection robot finds that there are branches in the pipeline during the inspection of the current buried pipeline, the inspection robot sets the pipeline mark bit pipeFlagNo_x_y for each branch pipeline, and then sends a request signal to the upper computer to request a response from the upper computer. After receiving the request signal, the upper computer tells the robot which branch pipeline to inspect based on the pipeline mark bit. The inspection robot regenerates the landmark node for the branch pipeline and inspects it normally. The above processing flow is as follows: Figure 6 .

[0164] 4) At the current landmark node, the inspection robot searches for all possible paths between the current node and the next landmark node, and calculates the distance of each path. The distance calculation method adopts the Mahalanobis distance, and the formula is as follows:

[0165]

[0166] 5) Sort the obtained distances, calculate the shortest path and the key path points passed by the path, and update the distance between the two landmark nodes.

[0167] 6) Determine whether the current underground landmark node number is the last node number in the landmark node sequence. If the current node number is equal to the last node number in the underground landmark node sequence, it means that the inspection robot has completed the inspection of the current underground pipeline, otherwise continue with step 4) and step 5).

[0168] 7) The inspection robot records the inspection results of the pipeline and uploads the inspection completion and pipeline flag information to the host computer.

[0169] Step 4: In the visual data processing module of the host computer, establish a simulation model of the inspection robot and view the relevant status operation data of the robot;

[0170] In the Xacro file, first add the robot base root tag base_link, and then in the root tag base_link, add various robot link components. In this system, the link components include a monocular industrial camera, a laser ranging radar, and a robotic arm. Set the shape and position of the link respectively. In this system, the industrial camera link camera is set to a square shape, located directly in front of the robot chassis. The laser ranging radar link lidar is cylindrical, located above the chassis and unobstructed. The robotic arm link marm is set to a multi-section rectangle, located at the tail of the chassis. Finally, set the joint connection mode of each link link to express the kinematic and dynamic characteristics of each component and the robot. The movement mode of the camera joint is fixed joint fixed, the connection mode of the radar link is revolute joint revolute, the connection mode of the robotic arm base link is revolute joint revolute, and the connection mode of the robotic arm upper link is floating joint floating.

[0171] Add components in Rviz to display the robot base and various sensor data to view the robot posture information and sensor information.

[0172] View the robot's motion status in the map in Gazebo.

[0173] Step 5: The inspection robot returns home or goes to the next inspection area after the inspection. After the inspection robot finishes the inspection of a pipeline in a certain area, it needs to go to the return point specified by the host computer or go to the next inspection area specified by the host computer.

[0174] When the inspection robot returns, it no longer needs to return according to the pipeline number, but directly plans the shortest path from the current point to the return point for return.

[0175] When the inspection robot goes to the next inspection area for inspection, it first directly plans the shortest path from the current point to the next area. After arriving at the inspection area, it checks whether there is an electronic map matching the current area in the map database. If there is no electronic map, the inspection robot will inspect along an edge along an S-shaped trajectory after arriving at the inspection area. After discovering the buried pipeline, it will report to the host computer to start the inspection. If there is an electronic map in the map database, the regional electronic map is directly retrieved, and according to the pipeline mark position given by the host computer, it goes to the starting end of the buried pipeline corresponding to the pipeline mark position, and inspects according to the numbering order in the pipeline mark node sequence.

Claims

1. An autonomous navigation system for a buried pipeline field inspection robot, characterized in that: include: Upper computer visualization data processing module, external environment multi-sensing module, multi-sensor information interaction storage module, bottom-level drive execution module, and electromagnetic holographic detection module; The external environment multi-sensing module is used to sense the ground surrounding environment information and generate surrounding environment image data and radar point cloud data; The electromagnetic holographic detection module is used to collect electromagnetic information of the secondary magnetic field induced by the buried pipeline and generate electromagnetic data; The electromagnetic holographic detection module includes an electromagnetic holographic detector and a mechanical arm; The mechanical arm is fixed on the cover plate of the crawler chassis of the inspection robot. The mechanical arm clamps a probe with an excitation coil to transmit electromagnetic signals to the underground pipeline. The electromagnetic holographic detector is installed on the crawler chassis of the inspection robot to collect electromagnetic data of the secondary magnetic field induced by the underground pipeline through the electromagnetic holographic detector. The bottom-level driving execution module is used to drive the crawler chassis of the inspection robot; The multi-sensor information interaction storage module is used to receive and upload surrounding environment image data, radar point cloud data, and electromagnetic data; The host computer visualization data processing module is used to construct a visualization map of the inspection area and a simulation model of the inspection robot, and generate a motion planning path for the inspection robot.

2. The autonomous navigation system for a buried pipeline field inspection robot according to claim 1, characterized in that: The external environment multi-sensing module includes a monocular industrial camera and a 3D laser radar; the multi-sensor information interaction storage module is a main control board chip, the monocular industrial camera is fixed on the track chassis cover of the inspection robot, and the 3D laser radar is fixed on the track chassis cover of the inspection robot through a radar rack; the monocular industrial camera and the 3D laser radar are electrically connected to the input end of the main control board chip respectively; the image data of the surrounding environment on the ground is collected by the monocular industrial camera, and the point cloud data of the surrounding obstacles is collected by the 3D laser radar.

3. The autonomous navigation system for a buried pipeline field inspection robot according to claim 1, characterized in that: The bottom-level driving execution module is a driving control board installed on the crawler chassis of the inspection robot, the control end of the driving control board is connected to the servo motor of the inspection robot, and the controlled end of the driving control board is connected to the main control board chip; The autonomous navigation system also includes a power supply module, which includes a ternary lithium battery pack and a battery voltage conversion module installed on the track chassis. The battery voltage conversion module converts the voltage output by the ternary lithium battery pack into the driving voltage required by each module.

4. An autonomous navigation method for a buried pipeline field inspection robot, implemented based on the autonomous navigation system for a buried pipeline field inspection robot according to any one of claims 1 to 3, characterized in that: The method comprises: Step 1: Start the field buried pipeline inspection robot and control the inspection robot to move to the base point of the area to be inspected; Step 2: Create a visual electronic map of the corresponding inspection area and initialize the electronic map; Step 3: Establish a motion planning strategy for the inspection robot based on the map information with the underground pipeline direction currently obtained by the inspection robot; Step 4: In the visual data processing module of the host computer, establish a simulation model of the inspection robot and view the relevant status operation data of the robot; Step 5: After the inspection, the inspection robot returns home or moves to the next inspection area.

5. The autonomous navigation method for a buried pipeline field inspection robot according to claim 4, characterized in that: The step 2 comprises: Step 2.1: Set the inspection area range parameters of the field inspection robot; Step 2.2: Start the inspection robot to build ground environment map and underground pipeline environment map models; Step 2.3: Generate a landmark node list according to the landmark nodes of the above-ground environment map and the underground pipeline environment map; the landmark node list includes: above-ground and underground landmark node numbers, timestamps corresponding to each landmark node, and distances between landmark nodes; Step 2.4: Using the timestamp of the landmark node generated by the inspection robot as the registration unit, match the above-ground and underground landmark nodes to obtain a set of landmark node sequences based on the underground pipeline landmark nodes; Step 2.5: Arrange the obtained landmark node sequence in ascending order of timestamps and store it in the multi-sensor information interaction storage module.

6. The autonomous navigation method for a buried pipeline field inspection robot according to claim 5, characterized in that: The process of establishing the ground environment map is as follows: Step A1: When the inspection robot moves to the inspection area, start the 3D laser radar and monocular industrial camera and set the time interval for the camera to collect images. The first image collected by the camera is used as the base point P of the inspection area coordinate system. world ; Step A2: The monocular industrial camera collects the feature point differences between different images at adjacent moments and calculates the camera's position P in the inspection area coordinate system through triangulation. camera ; Step A3: Based on the static coordinate relationship between the camera and the chassis of the inspection robot, obtain the posture equation x of the inspection robot in the inspection area coordinate system k =P1x k-1 +Q1u k +w k ; P1 is the posture transformation matrix of the inspection robot between adjacent moments, Q1 is the coordinate transformation coefficient, w k is the noise interference during the movement of the inspection robot; the inspection robot observation equation z is obtained k,j =P2y j +Q2x k +v k,j ; P2 is the measurement matrix of the image feature points, Q2 is the coefficient of the motion equation, v k,j The errors caused in the process of observing the ground environment; Step A4: Start the radar node and receive the three-dimensional point cloud data P of the inspection robot's surrounding environment and obstacles relative to the radar coordinate system lidar ; Step A5: According to the static coordinate relationship between the radar and the chassis of the inspection robot, the coordinates P of the surrounding targets in the inspection area coordinate system are obtained. object , used to optimize the inspection robot posture data calculated by the camera node; The construction process of the underground pipeline environment map model is as follows: During the movement of the inspection robot, an alternating current is applied to the coil on the probe to induce a primary magnetic field. The primary magnetic field induces a secondary magnetic field at the buried pipeline, and the secondary magnetic field data is received by the electromagnetic holographic detector. When the secondary magnetic field voltage reaches the minimum value, it means that the inspection robot has moved to the top of the buried pipeline. At this time, the landmark node is saved, and the buried pipeline observation equation is formed in combination with the current posture of the robot: k,i =(P3+d lift-off +r robot-arm )q i +Q3x k +e k,i , P3 is the observation matrix of pipeline road sign nodes, d lift-off is the vertical distance between the buried pipeline and the inspection robot, r robot-arm is the detectable area of ​​the inspection robot centered on the robotic arm, Q3 is the coefficient of the robot motion equation, e k,i To detect the errors in the underground pipeline environment.

7. The autonomous navigation method for a buried pipeline field inspection robot according to claim 5 or 6, characterized in that: When the obtained electronic map contains a ground environment map model and an underground pipeline environment map model, execute steps 2.2 to 2.5 to obtain a landmark node sequence in which the landmark nodes of the underground pipeline are sorted from small to large; When the obtained visualization map only has the ground environment map model but no underground pipeline environment map model, the inspection area range parameters set in step 1 are used as the standard, and the inspection area base point is used as the starting point in the inspection area to perform S-shaped curve movement along the length of the inspection area to detect underground pipelines. When underground pipelines are found, the landmark node sequence is generated in the order of small to large timestamps based on the current underground pipeline landmark node; When the obtained visualization map only contains the ground environment map model but no underground pipeline environment map model, and no underground pipeline is found while performing S-shaped curve movement along the length of the inspection area for underground pipeline detection, it is considered that there is no underground pipeline in this inspection area. The inspection robot stops moving and sends information to the upper computer to report that there is no underground pipeline here.

8. The autonomous navigation method for a buried pipeline field inspection robot according to claim 4, characterized in that: The step 3 comprises: Step 3.1: The inspection robot starts inspection from the node with the smallest landmark node number; Step 3.2: Search for all existing key path points between the current landmark node and the next landmark node and calculate the distance from the current landmark node to the next landmark node through the key path points; Step 3.3: Sort the obtained distances, calculate the shortest distance and its corresponding key path point, and update the distance information of the two landmark nodes in the landmark node list; Step 3.4: Control the inspection robot to move to the next landmark node at the shortest distance, and repeat step 3.3; Step 3.5: Update the distance between each landmark node in the corresponding landmark node list and record it.

9. The autonomous navigation method for a buried pipeline field inspection robot according to claim 4, characterized in that: When the inspection robot inspects the buried pipeline, the pipeline node numbering is processed as follows: ① The inspection robot covers the buried pipeline during the inspection: If the pipeline currently being inspected covers other surrounding pipelines, the inspection robot continues to inspect the pipeline normally according to the number in the landmark node sequence of the current pipeline; If the current inspection pipeline is covered by other surrounding pipelines, a pipeline road sign node is added at the starting position of the current inspection pipeline, and a pipeline node number is also added at the end position of the current inspection pipeline. At the same time, the coverage distance between the two nodes is calculated and saved. The current pipeline cannot be inspected within this coverage distance; ② When the inspection robot finds a branch in the pipeline during the inspection of the current buried pipeline, the inspection robot sets the pipeline mark bit pipeFlagNo_x_y for each branch pipeline, and then sends a request signal to the host computer to request a response from the host computer. After receiving the request signal, the host computer tells the robot which branch pipeline to inspect based on the pipeline mark bit. The inspection robot regenerates the landmark node for the branch pipeline and inspects it.