An intelligent control method for a buried pipeline detection robot

Abstract

The application belongs to the technical field of robot control, and specifically discloses an intelligent control method for a buried pipeline detection robot, which comprises the following steps: generating a global path based on a pipeline map and robot body parameters, controlling the robot to travel along the global path, collecting information through a front-looking sensor to identify a curve, determining a curve passing speed and a deceleration starting position based on curve geometric characteristics, body parameters and a current speed when the curve is identified, and planning a curve passing trajectory for each joint, comparing the actual trajectory of each joint with the curve passing trajectory in real time during the curve turning process, triggering local path re-planning when a deviation occurs, calculating an expected angle of each joint according to the real-time curvature of the re-planned path or the curve passing trajectory, and performing control; the application realizes stable detection in a complex pipeline environment through multi-joint trajectory safety planning and accurate corner control, as well as global path pre-planning, accurate curve determination and risk early warning.

Description

Technical Field

[0001] This invention belongs to the field of robot control technology and relates to an intelligent control method for a buried pipeline inspection robot. Background Technology

[0002] Buried pipelines are closed tubular structures buried underground in soil or rock strata using excavation or trenchless methods. Due to their long-term exposure to enclosed, narrow, humid environments prone to siltation, corrosion, and deformation, manual inspection faces challenges such as limited working space and high safety risks. Therefore, buried pipeline inspection robots have become crucial equipment for replacing manual labor in identifying internal pipeline defects and assessing their health. Consequently, achieving stable and adaptive movement of robots over bends or diameter changes, precise inspection based on posture and speed, preventing scraping of the pipe wall, and optimizing the travel route to improve inspection efficiency are key technical challenges that urgently need to be addressed in this field.

[0003] Currently, the control methods for underground pipeline inspection robots in the industry mainly rely on a conventional approach combining preset path tracking and sensor feedback correction. For example, Chinese invention patent CN114750165B discloses a precise positioning method and implementation device for an automated pipeline inspection robot. This scheme uses a magnetic induction intensity signal preliminary sampling circuit to collect magnetic induction intensity signals above the ground. After data processing, the signals are sent to a microprocessor analog-to-digital conversion module, combined with normal pipeline inspection positioning methods to determine the robot's position relative to the pipeline. When it is necessary to locate defects in the pipeline insulation layer, a steel rod is inserted into the ground to collect voltage signals. After preprocessing and separation circuits, the robot's position relative to the defect is determined using a defect location method. Finally, the robot is controlled to move directly above the defect, thereby improving the efficiency of automated pipeline inspection.

[0004] However, existing technologies based on preset paths and simple feedback have the following obvious limitations: First, the method is not adapted to the structure of multi-segment serial joint robots, focusing only on the overall positioning function. It does not combine joint width and pipe inner diameter for joint safety trajectory planning, nor does it provide a method for calculating the expected angle based on path feature points and combined with the rotation angle of the previous joint. As a result, it cannot meet the needs of dynamic posture control of multi-segment robots, and is prone to problems such as joint collision with pipe wall and adjacent joint rotation angle exceeding the limit, thereby reducing detection safety and trajectory stability.

[0005] Second, this method does not construct a complete control loop of global pre-planning, dynamic adjustment and local replanning. It only realizes the simple switching between normal line inspection and defect location. It cannot realize the overall path pre-planning based on pipeline map, the special control scheme for curves and the local replanning after trajectory deviation. Therefore, it cannot cope with complex curves and trajectory deviations, which can easily lead to problems such as loss of control of turning speed and path deviation, thus affecting the accuracy of detection data.

[0006] Therefore, there is an urgent need for a method that can achieve adaptive control of curves and complete closed-loop control in multi-segment serial articulated robot structures to solve the above-mentioned technical problems. Summary of the Invention

[0007] In view of this, in order to solve the problems mentioned in the background technology, an intelligent control method for underground pipeline inspection robots is proposed.

[0008] The objective of this invention can be achieved through the following technical solution: This invention provides an intelligent control method for a buried pipeline inspection robot, including: generating a global path based on a pipeline map and robot body parameters.

[0009] The robot is controlled to move along a global path and collects pipeline perception information in real time through forward-looking sensors, and then performs curve identification based on the pipeline perception information.

[0010] When the determination result is a curve, based on the curve geometry features, robot body parameters and current speed in the pipeline perception information, the robot's curve passage speed and deceleration starting position are determined, and the curve passage trajectory of each joint is planned according to the curve geometry features and robot body parameters.

[0011] During the robot's turning process, the actual trajectory of each joint is compared with the trajectory of the curve in real time. If a trajectory deviation occurs, local path replanning is triggered to generate a replanned path.

[0012] Based on the real-time curvature of the replanned path or curve trajectory, the desired angle of each joint is calculated, and corresponding control is executed.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention establishes kinematic constraint equations for each joint by combining the curvature radius of the bend, the inner diameter of the pipe and the robot body parameters, calculates the minimum safe distance between the trajectory and the pipe wall and adjusts the over-threshold trajectory, effectively maintains the safe distance between the joint and the pipe wall, avoids scratching and collision, ensures the robot can move safely in the closed narrow pipe, thereby improving the detection safety.

[0014] (2) This invention extracts the feature points of the local replanning path, calculates the relative rotation angle of the next joint relative to the previous joint, and verifies whether the relative rotation angle exceeds the preset limit, and finally determines the expected angle of each joint, so as to achieve precise control of the relative rotation angle of adjacent joints, avoid damage to the robot's mechanical structure caused by the joint rotation angle exceeding the limit, and at the same time ensure the coordination of multi-joint movement.

[0015] (3) This invention establishes a three-dimensional topology model by extracting features such as straight pipe sections from the underground pipeline map, and obtains the global path by verifying the path with robot joint parameters. At the same time, it achieves accurate judgment of curves through dual parameter comparison and secondary verification. Based on this, it adaptively plans and corrects the curve passage speed and joint passage trajectory, thereby effectively preventing operation failures caused by speed loss or path deviation and improving the passage ability in complex curve environments.

[0016] (4) This invention initiates secondary verification when the preliminary judgment result in the curve judgment is inconsistent with the global path, and outputs an early warning of the deviation between the overall path and the actual deviation or a detection deviation based on the verification result, thereby realizing timely identification and emergency handling of underground risks, reducing the risk of robot operation failure, and thus improving the reliability of the control method. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram showing the connections between the steps of the method of the present invention.

[0019] Figure 2 This is a schematic diagram illustrating the global path generation steps of the present invention.

[0020] Figure 3 This is a schematic diagram showing the connection steps of the curve-passing trajectory planning in this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] This invention first constructs a global path based on a priori pipeline map, taking into account the robot's kinematic constraints. During the robot's movement along the path, forward-looking sensors collect real-time information about the pipeline's shape and identify curves. When a curve is anticipated, the safe passage speed and deceleration starting position are calculated based on the curve's geometric features and the robot's current operating state. A curve-crossing trajectory that maintains a safe distance from the pipe wall is planned for each joint. During curve-crossing, the actual motion trajectory of each joint is monitored in real-time. If a deviation from the preset trajectory is detected, a local path replanning mechanism is triggered to generate a corrected path. Finally, based on the real-time curvature of the currently used path (replanned path or original curve trajectory), the desired rotation angle of each joint is calculated, and the corresponding actions are executed. Through this method, this invention effectively avoids the problems of collisions and the inability to autonomously correct trajectory deviations in multi-section robots in existing technologies, significantly improving the safety and control reliability of pipeline inspection robots in complex environments.

[0023] Please see Figure 1 As shown, the present invention provides an intelligent control method for a buried pipeline inspection robot, comprising the following steps S1 to S5.

[0024] S1. Generate a global path based on the pipeline map and robot body parameters.

[0025] Please see Figure 2 As shown, this step aims to pre-plan a feasible path from the starting point to the ending point based on the prior information of the pipeline and the robot's own motion capabilities, providing a basic reference for subsequent real-time control. Specifically, it includes the following sub-steps: S1-1, Constructing a three-dimensional topology model of the pipeline.

[0026] The geometric features and topological connections of straight pipe sections, bends, diameter changes, and branch nodes are extracted from the underground pipeline map. The geometric features include the length, diameter, bending radius, diameter change location, and angle of each pipe section; the topological connections clarify the sequence of connections between pipe sections and the direction of branching. Based on this information, a structured 3D pipeline topology model is constructed. This model not only includes the spatial geometric outline of the pipeline but also stores the precise coordinates and connection attributes of each key node, providing complete spatial data support for subsequent route planning.

[0027] S1-2, Generate the initial path set.

[0028] Using the robot's starting position as the initial point and the detection endpoint as the target point, a path search algorithm (such as the A algorithm or Dijkatra's algorithm) is used to traverse all possible paths in the three-dimensional pipeline topology model to generate several initial paths from the starting point to the endpoint. Each initial path consists of a series of continuous path units, which correspond to straight pipe sections, bends, diameter-changing sections, or branch nodes in the topology model.

[0029] S1-3. Perform kinematic feasibility verification on the initial path.

[0030] The minimum turning radius of each joint and the maximum allowable relative rotation angle between adjacent joints are extracted from the robot body parameters. For each initial path, it is divided into continuous path units, and the key geometric parameters of each unit are labeled: if it is a curved pipe segment, the actual radius of curvature is extracted; if it is a straight pipe segment, the axial length is labeled; at the same time, the angle between two adjacent path units is calculated, that is, the angle of change of the path direction. The following two verifications are performed: (1) Minimum turning radius verification: the actual radius of curvature of each curved pipe segment is compared with the minimum turning radius of each joint of the robot one by one. If the actual radius of all curved pipe segments is greater than or equal to the minimum turning radius of the corresponding joint, the path unit is determined to satisfy the minimum turning radius constraint; otherwise, the path unit is determined to be infeasible.

[0031] (2) Verification of adjacent joint rotation angles: Based on the included angle between adjacent path units, combined with the total number of robot joints and the kinematic coupling relationship of each joint, determine the theoretical relative rotation angle required for each joint when the robot passes through the connection point. The theoretical relative rotation angle is characterized by the angle of deflection required by each joint relative to the previous joint when the overall steering of the robot matches the change in path direction.

[0032] As an example of allocation, when the robot adopts a uniform steering strategy, the theoretical relative rotation angle of each joint can be set to be equal, that is, equal to the angle between adjacent path units divided by the total number of robot joints; when the robot adopts a head joint-dominated steering strategy, the theoretical relative rotation angle of the head joint is equal to the angle between adjacent path units, and the theoretical relative rotation angle of subsequent joints is zero.

[0033] The maximum permissible relative rotation angle between adjacent joints is obtained from the robot's body parameters. This parameter is a pre-calibrated upper limit of the relative rotation angle between joints, used to avoid mechanical interference or structural damage caused by excessive rotation angle.

[0034] The calculated theoretical relative rotation angles of each joint are compared with the maximum allowable relative rotation angles of their respective adjacent joints. If the theoretical relative rotation angles of all joints do not exceed their maximum allowable relative rotation angles, the path element is determined to satisfy the adjacent joint rotation angle constraints. If the theoretical relative rotation angle of any joint exceeds its maximum allowable relative rotation angle, the path element is determined not to satisfy the adjacent joint rotation angle constraints.

[0035] If all path units of an initial path simultaneously satisfy both of the above constraints, then the initial path is marked as a feasible initial path; otherwise, it is marked as an infeasible initial path. Traverse all initial paths to obtain the set of feasible initial paths.

[0036] S1-4. Select the global path.

[0037] Calculate the total length of each feasible initial path, which is the sum of the axial lengths of each path element, and select the feasible initial path with the shortest path length as the final output global path. This global path is the optimal reference path that satisfies the robot's motion capabilities without considering real-time environmental changes.

[0038] S2. Control the robot to move along the global path and collect pipeline perception information in real time through the forward-looking sensor, and identify curves based on the pipeline perception information.

[0039] This step involves using forward-looking sensors (such as LiDAR and structured light cameras) to perceive the shape of the pipeline ahead in real time during the robot's actual movement. This perception is then fused with global path information to accurately determine whether there is a curve ahead, providing a reliable basis for subsequent curve control. Specifically, it includes the following sub-steps: S2-1, Real-time acquisition of pipeline perception information.

[0040] As the robot travels along the global path, the forward-looking sensor emits detection signals forward at a fixed frequency. After receiving the reflected signals, it generates three-dimensional point cloud data or depth images containing the contour of the pipe's inner wall, serving as pipe perception information. This information includes at least the spatial coordinates, cross-sectional shape, and axial extension direction of each detection point in the pipe ahead.

[0041] S2-2, Calculate the characteristic parameters of the curve.

[0042] The following two bend characteristic parameters are extracted from the pipeline sensing information for each detection point: S2-2-1, pipeline cross-section ellipticity: An ellipse is fitted to the pipeline cross-section at each detection point, and the ratio of the major axis to the minor axis is calculated to obtain the cross-sectional ellipticity at each detection point. Ellipticity reflects the degree of radial deformation of the pipeline.

[0043] S2-2-2, Pipeline Axial Profile Curvature: Along the pipeline axis, curve fitting is performed on the coordinates of the center points of multiple consecutive detection points. The curvature value of the fitted curve at each detection point is calculated to obtain the axial profile curvature. This curvature reflects the degree of bending of the pipeline axis.

[0044] S2-3, Preliminary determination of the existence of the curve.

[0045] The ellipticity and axial profile curvature of each detection point are compared with preset ellipticity thresholds and preset curvature thresholds, respectively. The proportion of detection points whose ellipticity and curvature both exceed the corresponding thresholds to the total number of detection points in the area ahead of that segment is defined as the overall compliance rate of the curve feature.

[0046] If the overall compliance rate of curve features is greater than the preset compliance rate threshold, such as 80%, then it is preliminarily determined that there is a curve ahead; otherwise, it is preliminarily determined that there is no curve ahead.

[0047] The preset ellipticity threshold is determined based on the pipe material, nominal diameter, and industry standards (such as ISO 11922), and a safety redundancy is reserved for the robot joint width. The preset curvature threshold is calculated based on the minimum bend radius specified in the pipe design specification (such as GB 50253) and corrected by combining the measured data of the minimum turning radius of the robot joint. The preset compliance rate threshold is set according to the sensor accuracy and the degree of environmental interference, and is usually taken as 70% to 90%.

[0048] S2-4. Consistency verification and secondary verification.

[0049] The preliminary judgment result is compared with the corresponding road segment type (curve or straight pipe) marked in the global path: if the two are consistent, that is, the preliminary judgment is that there is a curve and the global path is marked as a curve, or the preliminary judgment is that there is no curve and the global path is marked as a straight pipe, then the preliminary judgment result is used as the final curve identification result.

[0050] If the two results are inconsistent, a secondary verification process is initiated: First, the forward-looking sensor undergoes self-calibration (e.g., adjusting gain and filtering parameters). Then, multiple cycles of forward-looking data are continuously collected, and the overall compliance rate of the curve features is recalculated. If the secondary verification result is still consistent with the initial judgment, the initial judgment result is used as the final identification result, and a global path deviation warning is issued, indicating that the global path may be outdated or contain errors. If the secondary verification result is consistent with the global path labeling, the global path labeling is used as the final identification result, and a detection deviation warning is issued, indicating that the sensor may have continuous interference or malfunction.

[0051] This cross-validation mechanism effectively combines the advantages of prior maps and real-time perception, avoiding control errors caused by misjudgments from a single information source.

[0052] S3. When the determination result is a curve, based on the curve geometry features, robot body parameters and current speed in the pipeline perception information, determine the robot's curve passage speed and deceleration starting position, and plan the curve passage trajectory of each joint according to the curve geometry features and robot body parameters.

[0053] Please see Figure 3 As shown, this step is the core preparation stage for navigating the curve. By calculating the safe speed, optimal deceleration point, and joint trajectory, it ensures that the robot smoothly and safely enters and passes through the curve. Specifically, it includes the following sub-steps: S3-1, Determine the curve passage speed.

[0054] First, based on the bend curvature radius in the pipeline sensing information. and the physical length of each joint in the robot's body parameters and the maximum permissible angular velocity of the joint Calculate the maximum linear velocity of the robot based on joint angular velocity constraints. , In the formula, Indicates the joint number. , This indicates taking the minimum value of the product across all joints.

[0055] Secondly, the friction coefficient between the inner wall of the pipe and the contact surface between the robot and the robot is extracted from the robot's body parameters. and the radius of curvature of the curve The computer body does not slide at its maximum speed , In the formula, For gravitational acceleration, from and The minimum value is selected as the safe passage speed threshold.

[0056] Finally, the robot's current speed is compared with the safe passage speed threshold: if the current speed is less than or equal to the safe passage speed threshold, the current speed is used as the curve passage speed; otherwise, the safe passage speed threshold is used as the curve passage speed, meaning the robot needs to decelerate to that speed.

[0057] S3-2. Determine the location of the deceleration point.

[0058] First, based on the reference deceleration in the robot's body parameters. (Usually, the normal braking deceleration is taken, combined with the current speed) Speed ​​of travel on curves Calculate the minimum deceleration distance required to reduce from the current speed to the cornering speed. : .

[0059] Secondly, the location of the bend's starting point is identified using pipeline sensing information, and combined with the robot's built-in odometer information, the estimated pipeline axis distance from the robot's current position to the bend's starting point is calculated.

[0060] Then, the minimum deceleration distance is compared with the estimated pipe axis distance: if the minimum deceleration distance is less than or equal to the estimated pipe axis distance, the robot is offset along the pipe axis from the bend start point towards its current position by the minimum deceleration distance, and the offset position is determined as the deceleration start point. At this point, the robot starts to decelerate at the reference deceleration rate, and when it reaches the bend start point, its speed just drops to the bend travel speed.

[0061] If the minimum deceleration distance is greater than the estimated distance along the pipeline axis, it means that the remaining distance is insufficient to complete a smooth deceleration. At this time, the current position is immediately taken as the deceleration starting point, and emergency braking is triggered, that is, the maximum deceleration is activated, and if necessary, the joint locking mechanism is activated to minimize the speed when entering the curve and reduce the safety risk.

[0062] S3-3, Plan the curve trajectory of each joint.

[0063] This step plans a desired trajectory for each joint of the robot, starting from its current position, smoothly navigating the curve without colliding with the pipe wall.

[0064] S3-3-1. Determine spatial constraints: Based on the curvature radius of the curve, the inner diameter of the pipe, and the robot body parameters, clarify the spatial constraints for the movement of each joint, including: the maximum turning angle of a single joint, the minimum distance between adjacent joints, and the minimum safe distance between the joint and the pipe wall, which are determined by subtracting half of the robot's maximum radial dimension from the inner diameter of the pipe.

[0065] S3-3-2. Planning the head joint entry trajectory segment: Using the tangent direction of the pipeline centerline as the tangent direction of the head joint trajectory segment, the head joint enters the curve from its current position along this tangent direction. This entry segment trajectory is a curve tangent to the pipeline centerline in the horizontal plane and remains coaxial with the pipeline in the vertical plane.

[0066] S3-3-3, Generating Subsequent Joint Following Trajectory Segments: A joint position coupling algorithm is used to sequentially generate the following trajectories for each subsequent joint based on the spatial constraints of adjacent joints. This algorithm uses the planned trajectory of the previous joint as a reference, and by constraining the relative positions and angles of adjacent joints at the hinges, ensures that the starting point of the trajectory of the subsequent joint smoothly connects to the ending point of the trajectory of the previous joint at the hinge, and that the angle between adjacent trajectory segments does not exceed the maximum steering angle of the joint. In practical implementation, an inverse kinematics solution method based on spline interpolation can be used to represent the joint position as a function of time.

[0067] S3-3-4. Trajectory Safety Adjustment: For the initial trajectory of each joint, calculate the real-time distance between each point on the trajectory segment and the pipe wall based on the joint width and the pipe inner diameter. If the real-time distance is less than the minimum safety distance, shift the trajectory at that point radially inward until the safety distance requirement is met. The offset is smoothly transitioned using linear or nonlinear interpolation to avoid introducing new discontinuities.

[0068] S3-3-5. Integrating and Generating the Final Trajectory: The adjusted joint trajectories are integrated according to the time sequence to obtain the complete curve passage trajectory. This trajectory provides a sequence of the expected position changes over time for each joint, which is the basis for subsequent trajectory comparison and replanning.

[0069] S4. During the robot's turning process, the actual trajectory of each joint is detected and compared with the trajectory of the curve in real time. If a trajectory deviation occurs, local path replanning is triggered to generate a replanned path.

[0070] This step is responsible for monitoring the robot's actual movement status in real time during curve passage. Once a deviation from the predetermined trajectory is detected, dynamic correction is immediately initiated to ensure the robot always passes safely. Specifically, it includes the following sub-steps: S4-1, Real-time trajectory comparison.

[0071] Align the actual motion trajectories of each joint with the curve passing trajectory planned in S3-3 on the same time axis, and establish a trajectory comparison coordinate system. Usually, the center line of the pipeline is used as the reference, with the direction of pipeline travel as the x-axis, the radial direction as the y-axis, and the vertical direction as the z-axis.

[0072] At each sampling moment, the Euclidean distance between the actual position of each joint and the corresponding position on the curve trajectory is calculated as the instantaneous position deviation value. At the same time, the mean and maximum position deviations of all sampling points within a sliding window (e.g., the most recent second) are calculated.

[0073] The average position deviation is compared with a preset deviation threshold, and it is also determined whether the maximum deviation value exceeds the emergency deviation threshold.

[0074] If the average position deviation is less than the preset deviation threshold and the maximum deviation is less than or equal to the emergency deviation threshold, it is determined that the trajectory has not deviated significantly, and the original trajectory continues to be executed.

[0075] If the average position deviation is greater than or equal to the preset deviation threshold or the maximum deviation is greater than the emergency deviation threshold, the trajectory is determined to have deviated, and the process proceeds to the next step of local path replanning.

[0076] The preset deviation threshold is set according to the robot control accuracy requirements and sensor noise level, and is usually 5% to 10% of the joint width; the emergency deviation threshold is set according to the pipe wall safety distance, for example, half of the minimum safety distance.

[0077] Furthermore, to avoid system oscillations caused by frequent replanning triggers, this method introduces a replanning cooling mechanism. Specifically, after each local path replanning is completed, the system initiates a cooling window of a preset duration (e.g., 1-2 control cycles or 0.5 seconds). Within this window, even if a trajectory deviation is detected, only an alarm is recorded, and no new replanning is triggered to maintain stable tracking of the current path. After the cooling window ends, if the trajectory deviation persists, the ability to trigger replanning is restored.

[0078] S4-2, Local path replanning.

[0079] When a trajectory deviation is detected, local path replanning is immediately triggered to correct the robot's motion. The replanning process is as follows: S4-2-1, Determine the initial state for replanning: The actual pose (position and attitude) and velocity of each joint at the current moment are used as the initial state and the starting point for the replanning path.

[0080] S4-2-2, Delineate the replanning space: Based on the 3D topology model of the pipeline and the robot's current position, a local pipeline region extending forward along the pipeline axis from the current position for a predetermined length is selected as the replanning space. This region contains the geometric information of subsequent pipelines, ensuring that the replanned path can connect with the trajectory of the curves that have not deviated ahead.

[0081] S4-2-3. Generating Local Candidate Paths: Within the replanning space, using kinematic constraints of each joint (minimum turning radius, maximum turning angle, etc.) as conditions, an optimization algorithm (such as Fast Expanding Random Tree (RRT) or Model Predictive Control (MPC)) is employed to generate a series of local candidate paths from the current position to the target point. The selection of the target point must ensure that the path length is minimized and the path is smooth.

[0082] S4-2-4. Feasibility Verification: A rigorous feasibility verification is performed on each candidate local path, including: Corner Constraint Verification: The joint poses of keyframes (e.g., equally spaced sampling points) are extracted, the relative rotation angles of adjacent joints are calculated, and compared with the maximum permissible relative rotation angle threshold for adjacent joints in the robot's body parameters. If the relative rotation angles of all keyframes do not exceed the threshold, the verification passes; otherwise, it fails.

[0083] Distance constraint verification: Calculate the minimum real-time distance between each joint on the path and the pipe wall, and compare it with a preset safe distance threshold. If all real-time distances are greater than or equal to the safe distance threshold, the verification passes; otherwise, it fails.

[0084] If a candidate path satisfies both constraints, it is marked as a feasible local path; otherwise, the path is iteratively adjusted based on the deviation value of the failed constraint. For example, when the distance constraint fails, the value of the deviation from the safe distance is used as a penalty term, and the weight of the penalty term is increased in the cost function of the replanning algorithm to regenerate the local path; when the corner constraint fails, the path curve is smoothed to reduce its rate of curvature change until the verification is passed or the maximum number of iterations (e.g., 10 times) is reached.

[0085] S4-2-5. Generate a replanning path: Smoothly connect the validated local path with the subsequent curve-passing trajectory that has not deviated (e.g., using fifth-order polynomial interpolation) to generate a complete replanning path. This path starts from the current position, goes through a correction segment, and then smoothly returns to the original curve-passing trajectory.

[0086] If a feasible local path cannot be generated even after reaching the maximum number of iterations, an infeasibility warning will be output, and a global path correction mechanism will be triggered, such as stopping the movement and uploading the status, requesting remote instructions, or reverting to a safe position.

[0087] S5. Calculate the desired angle of each joint based on the real-time curvature of the replanned path or curve trajectory, and execute the corresponding control.

[0088] This step converts path information into joint execution commands, enabling precise angle control to ensure the robot's actual movement conforms to the desired trajectory. Specifically, it includes the following sub-steps: S5-1, Extracting path feature points.

[0089] Feature points are extracted from the currently adopted target path (the curve trajectory if no replanning is triggered, and the replanned path if replanning has been triggered) at a certain step size. These feature points include the path's start point, end point, inflection points where curvature changes significantly, points of curvature change, and safety constraint points near the pipe wall. Each feature point records its spatial coordinates and tangent direction, i.e., the orientation of the path at that point.

[0090] S5-2, Calculate the desired angle of the head joint.

[0091] The robot's current pose (obtained by fusion of odometry and attitude sensors) is used as the starting point for the head joint. The angle between the direction of the line connecting the starting point to the first feature point and the reference direction of the pipeline centerline (i.e., the tangent direction of the global path at the current position) is taken as the desired rotation angle of the head joint. This rotation angle reflects the angle by which the head joint needs to deflect to align with the starting direction of the path.

[0092] S5-3, Calculate the expected angle of the subsequent joint.

[0093] For the second joint and each subsequent joint, the expected angle is calculated sequentially using a recursive method: S5-3-1, Calculate the relative rotation angle of the next joint: The angle between the vector from the starting position of the next joint to the current feature point and the vector from the previous joint to the next joint is used as the relative rotation angle of the next joint.

[0094] S5-3-2. Calculate the preliminary expected angle: Sum the expected angle of the previous joint with the relative angle of the next joint to obtain the preliminary expected angle of the next joint.

[0095] S5-3-3, Angle Limiting Verification: Calculate the difference between the initial expected angle of the subsequent joint and the expected rotation angle of the preceding joint, i.e., the actual relative rotation angle, and determine whether the absolute value of this difference exceeds a preset rotation angle difference threshold. This threshold is set based on the coordinated movement capability of adjacent joints to avoid mechanical interference caused by excessive relative rotation angles. If the absolute value of the actual relative rotation angle does not exceed the preset threshold, the initial expected angle is used as the expected angle of the subsequent joint; if it does exceed the threshold, the initial expected angle is limited so that the actual relative rotation angle equals the preset threshold, and the direction is the same as the relative rotation angle. That is, the expected angle of the subsequent joint equals the expected rotation angle of the preceding joint plus the preset rotation angle difference threshold with a positive or negative sign.

[0096] S5-3-4, Single Joint Limiting Verification: Further check whether the expected angle of the next joint obtained after the above limiting exceeds the maximum permissible rotation angle of the joint itself. If it exceeds, then the maximum permissible rotation angle of the joint itself is taken as the expected angle of the next joint.

[0097] Repeat the above process to calculate the desired angles of all joints in turn.

[0098] S5-4, Execution Control.

[0099] The calculated desired angles of each joint are converted into control commands and sent to the joint servo drivers to drive the robot to move along the desired path. At the same time, the control system provides real-time feedback on the actual joint angles in a high-speed closed-loop manner and repeats the trajectory comparison in S4-1 to form a complete closed loop.

[0100] Through the above steps S1 to S5, the present invention realizes the intelligent adaptive passage of the underground pipeline inspection robot in complex pipeline environments, and improves the curve-passing ability, trajectory tracking accuracy and operational safety of the multi-joint robot.

[0101] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0102] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0103] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0104] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0105] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent control method for a buried pipeline inspection robot, characterized in that: The method includes: A global path is generated based on the pipeline map and robot body parameters. The robot is controlled to move along a global path and collects pipeline perception information in real time through a forward-looking sensor, and then performs curve identification based on the pipeline perception information. When the determination result is a curve, based on the curve geometry features, robot body parameters and current speed in the pipeline perception information, the robot's curve passage speed and deceleration starting position are determined, and the curve passage trajectory of each joint is planned according to the curve geometry features and robot body parameters. During the robot's turning process, the actual trajectory of each joint is detected and compared with the curve's trajectory in real time. If a trajectory deviation occurs, local path replanning is triggered to generate a replanned path. Based on the real-time curvature of the replanned path or curve trajectory, calculate the desired angle of each joint and execute the corresponding control. The calculation of the desired angles for each joint includes: Based on the real-time curvature of the path or curve trajectory through local replanning, the position coordinates and tangent direction of each feature point on the path are extracted. Using the robot's current pose as the starting point of the head joint, the angle between the direction of the line connecting the starting point to the feature point and the reference direction of the pipeline centerline is taken as the desired rotation angle of the head joint. Calculate the angle between the vector from the starting position of the next joint to the feature point and the vector from the previous joint to the next joint, and use it as the relative rotation angle of the next joint. The relative rotation angle is summed with the expected rotation angle of the preceding joint to obtain the preliminary expected angle of the following joint. If the difference between the initial expected angle and the expected angle of the previous joint is greater than a preset angle difference threshold, the initial expected angle is limited to obtain the expected angle of the next joint; otherwise, the initial expected angle is used as the expected angle of the next joint, and the expected angles of each joint are obtained accordingly.

2. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: The generated global path includes: Straight pipe sections, bends, diameter-changing sections, and branch nodes are extracted from the underground pipeline map, and a three-dimensional pipeline topology model is established accordingly. Using the robot's starting position as the initial point and the detection endpoint as the target point, initial paths are generated through a path planning algorithm. The minimum turning radius of each joint and the maximum relative rotation angle of adjacent joints are extracted from the robot body parameters. The feasibility of the initial path is verified to obtain each feasible initial path. Based on the path length of each feasible initial path, the shortest feasible initial path is selected as the global path.

3. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: The curve identification process includes: The pipe cross-sectional ellipticity and axial profile curvature of each detection point in the pipeline sensing information are obtained and compared with the corresponding preset thresholds. The percentage of detection points where both ellipticity and axial curvature exceed the threshold is used as the overall compliance rate of curve features. If the overall compliance rate of the curve feature is greater than its threshold, it is preliminarily determined that there is a curve ahead; otherwise, it is preliminarily determined that there is no curve ahead. The preliminary judgment result is compared with the global path. If they match, the preliminary judgment result is taken as the curve judgment result. If there is a discrepancy, a second verification is initiated. If the result of the second verification is consistent with the initial judgment, the initial judgment result is used as the curve judgment result, and a path deviation warning is issued. Otherwise, the global path label is used as the curve judgment result, and a detection deviation warning is issued.

4. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: The determination of the robot's cornering speed includes: Based on the curvature radius of the bend in the pipeline sensing information and the physical length and maximum allowable angular velocity of the joint in the robot body parameters, the maximum linear velocity of the robot based on the joint angular velocity constraint is calculated. Based on the coefficient of friction and the radius of curvature of the curve, the maximum speed at which the computer body does not sideslip is calculated, and the minimum value between the maximum linear velocity of the joint and the maximum speed at which the computer body does not sideslip is selected as the safe passage speed threshold. The robot's current speed is compared with the safe passage speed threshold. If the current speed is less than or equal to the safe passage speed threshold, the current speed is used as the curve passage speed; otherwise, the safe passage speed threshold is used as the curve passage speed.

5. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: Determining the robot's deceleration starting point position includes: Based on the reference deceleration in the robot's body parameters, and combined with the current speed and the cornering speed, calculate the minimum deceleration distance required to reduce from the current speed to the cornering speed; Based on pipeline sensing information, the location of the bend's starting point is identified. Combined with the robot's built-in odometer information, the estimated pipeline axis distance from the robot's current position to the bend's starting point is calculated and compared with the minimum deceleration distance. If the distance along the pipe axis is greater than or equal to the minimum deceleration distance, the robot will deflect in the opposite direction from the start of the bend toward the current position along the pipe axis by the minimum deceleration distance, and the position after the deflection will be used as the starting position of deceleration. Otherwise, the current position will be used as the starting position of deceleration, and emergency braking will be triggered.

6. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: The planned curve-passing trajectory for each joint includes: Based on the curvature radius of the curve, the inner diameter of the pipe, and the robot body parameters, the spatial constraints for the movement of each joint are determined. The tangent direction of the pipeline centerline is used as the tangent direction of the trajectory segment, and the curve entry trajectory segment of the head joint is planned accordingly. Based on the spatial constraints of adjacent joints, a joint position coupling algorithm is used to generate the following trajectory segments for each joint; Based on the width of each joint and the inner diameter of the pipe, the real-time distance between the trajectory segment and the pipe wall is calculated. The trajectory segments that exceed the minimum safe distance are offset inward and adjusted. The trajectory segments are then integrated according to the time sequence to generate the curve passing trajectory of each joint.

7. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: The real-time comparison includes: Align the actual motion trajectories of each joint with the curve passage trajectory on the same time axis to establish a trajectory comparison coordinate system; Calculate the Euclidean distance between the actual position of each joint and the corresponding position of the curve trajectory at the same instant to obtain the instantaneous position deviation value, and calculate the mean and maximum position deviation. The average position deviation is compared with a preset deviation threshold, and it is determined whether the maximum deviation value exceeds the emergency deviation threshold. If the average position deviation is less than the preset deviation threshold and the maximum deviation does not exceed the emergency deviation threshold, the trajectory is determined to be without deviation; otherwise, the trajectory is determined to be deviated.

8. The intelligent control method for a buried pipeline inspection robot according to claim 1, characterized in that: The triggering of local path replanning includes: The initial state is planned with respect to the current pose and velocity of each joint. Based on the 3D topology model of the pipeline and the current position of the robot, a local pipeline area extending forward along the pipeline axis from the current position to a predetermined length is selected as the replanning space. Within the replanning space, a local path from the current position to the curve passage trajectory is regenerated, conditioned on the kinematic constraints of each joint. The feasibility of the replanned local path is verified, and the verified local path is smoothly connected with the subsequent curves through the trajectory to generate the local replanned path.

9. The intelligent control method for a buried pipeline inspection robot according to claim 8, characterized in that: The feasibility verification includes: Extract joint pose data from local path keyframes, calculate the relative rotation angle of adjacent joints, and compare it with the maximum allowable relative rotation angle threshold of adjacent joints in the robot body parameters; If the relative rotation angles of adjacent joints in all keyframes do not exceed the maximum allowable threshold, the rotation constraint verification is deemed to have passed; otherwise, the rotation constraint verification is deemed to have failed. Extract the minimum real-time distance between each joint trajectory and the pipe wall from the local path, and compare it with the preset safe distance threshold; If the real-time distances are all greater than or equal to the preset safe distance threshold, the distance constraint verification is deemed to have passed; otherwise, the distance constraint verification is deemed to have failed. If both the corner constraint and distance constraint verifications pass, the local replanning path is deemed feasible. If any verification fails, the local path is iteratively adjusted based on the deviation value of the failed item, and re-verified until all verifications pass or the maximum number of iterations is reached; If the maximum number of iterations is reached and the path still fails, an infeasibility warning will be output, triggering the global path correction mechanism.

Images