A pipeline robot navigation method and system

Abstract

The application discloses a pipeline robot navigation method and system, and relates to the technical field of robot navigation control; wherein the method comprises the following steps: in response to an obstacle avoidance instruction, target point attraction and obstacle repulsion are calculated by using an artificial potential field method according to target point position and obstacle information; based on the attraction and the repulsion, a driving direction and a target included angle of an adjacent joint unit are determined; the pipeline robot is adjusted from a first posture to a second posture according to the driving direction and the target included angle to avoid obstacles; the first posture or the second posture is at least one of a straight-line posture, an S-shaped posture, an arc-shaped posture, a spiral posture and a wave-shaped posture; and the first posture is different from the second posture. The pipeline robot can efficiently avoid obstacles in a complex pipeline environment through the cooperation of the artificial potential field method and the multi-joint unit posture adjustment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot navigation control, in particular to a pipeline robot navigation method and system. BACKGROUND

[0002] In the oil, gas, chemical and other industries, pipeline transportation is the core way of material transportation, and the complexity of the internal structure of the pipeline poses a severe challenge to detection and maintenance. The pipeline robot needs to work in various structures such as straight pipes, elbow pipes, tees and reducing pipes, and also needs to deal with static obstacles (such as weld steps and stones) and dynamic obstacles (such as flowing sediments).

[0003] In the prior art, the information interaction mechanism of global path planning and local path planning has significant defects. The global planning cannot effectively pass the details such as narrow pipe sections and obstacle dense areas to the local planning, resulting in that the local planning lacks environmental detail support and it is difficult to generate an efficient obstacle avoidance path. In addition, the traditional obstacle avoidance method does not fully consider the mechanical configuration characteristics of the pipeline robot, for example, the robot cannot flexibly adjust the body posture in a closed pipeline, which leads to improper handling of repulsive force and the robot is easily blocked by local obstacles. These problems seriously restrict the autonomy and adaptability of the robot in complex pipeline environments and reduce the detection and maintenance efficiency. SUMMARY

[0004] To solve the problems in the prior art, the present application provides a pipeline robot motion control method and system, which can make the pipeline robot efficiently avoid obstacles in a complex pipeline environment through the cooperation of artificial potential field method and multi-joint unit posture adjustment.

[0005] To achieve the above-mentioned purpose, the present application provides a pipeline robot navigation method, the pipeline robot comprising a plurality of actively connected joint units, the method comprising:

[0006] In response to an obstacle avoidance instruction, the target point attraction force and the obstacle repulsion force are calculated by using the artificial potential field method according to the target point position and the obstacle information;

[0007] Based on the attraction force and the repulsion force, the driving direction and the target included angle of the adjacent joint unit are determined;

[0008] According to the driving direction and the target included angle, the pipeline robot is adjusted from a first posture to a second posture to avoid obstacles; the first posture or the second posture is at least one of a straight posture, an S-shaped posture, an arc-shaped posture, a spiral posture and a wave-shaped posture; the first posture is different from the second posture.

[0009] Optionally, the target point position is from a remote end, and the method further comprises:

[0010] Based on the loss value of the predefined pipeline structure feature and the target point position, a total loss function of each candidate path in the pipeline global map is calculated by a hybrid A-star algorithm, and a global path is determined by minimizing the total loss function.

[0011] Optionally, the method further comprises:

[0012] When the pipeline robot successfully avoids the obstacle, obstacle distribution data and an effective path segment are generated according to the corresponding obstacle information, the driving direction and the target angle;

[0013] Based on the obstacle distribution data and the effective path segment, the pipeline global map is updated.

[0014] Optionally, the pipeline robot further comprises an internal sensor module and an external sensor module; and the method for generating the obstacle avoidance instruction comprises:

[0015] Obtaining state data of the pipeline robot body collected by the internal sensor module and pipeline environment data collected by the external sensor module;

[0016] Based on the pipeline environment data, obstacle information is determined;

[0017] According to the state data and the obstacle information, an obstacle avoidance instruction is generated.

[0018] Optionally, when the obstacle information indicates that the obstacle is a dynamic obstacle, an artificial potential field method is used to calculate the obstacle repulsion force, comprising:

[0019] Based on the obstacle information, a motion prediction model pre-trained is used to generate a motion trajectory of the dynamic obstacle;

[0020] According to the motion trajectory, the repulsion force calculation parameter in the artificial potential field method is adjusted to dynamically calculate the obstacle repulsion force.

[0021] Optionally, the first attitude is a straight-line attitude, and the second attitude is a spiral attitude; and according to the driving direction and the target angle, the pipeline robot is adjusted from the first attitude to the second attitude to avoid the obstacle, comprising:

[0022] According to the driving direction and the target angle, each joint unit is controlled to rotate cooperatively around the pipeline axis direction, so that the pipeline robot is converted from the straight-line attitude to the spiral attitude, forming a spiral-shaped advancing path to bypass the obstacle.

[0023] Optionally, the calculation formula of the attractive force is:

[0024] When d goal d0, F att =k att xdgoal ;

[0025] When d goal ≥ d0,

[0026] In the formula, F att is an attractive force, k att is an attractive force coefficient, d goal is the distance from the pipeline robot to the target point, and d0 is an attractive force distance threshold value;

[0027] The calculation formula of the repulsive force is as follows:

[0028] When d obs < d1,

[0029] When d obs > d1, F rep = 0;

[0030] In the formula, F rep is a repulsive force, k rep is a repulsive force coefficient, d obs is the distance from the pipeline robot to the obstacle, d1 is a repulsive force distance threshold value, is a unit vector from the obstacle to the robot.

[0031] The application further provides a pipeline robot navigation system, the pipeline robot comprising a plurality of actively connected joint units, and the system comprising:

[0032] An obstacle avoidance decision unit is configured to, in response to an obstacle avoidance instruction, calculate a target point attractive force and an obstacle repulsive force by using an artificial potential field method according to a target point position and obstacle information;

[0033] A motion planning unit is configured to determine a driving direction and a target included angle of an adjacent joint unit based on the attractive force and the repulsive force;

[0034] A posture control unit is configured to adjust the pipeline robot from a first posture to a second posture according to the driving direction and the target included angle, so as to avoid the obstacle; the first posture or the second posture is at least one of a straight-line posture, an S-shaped posture, an arc-shaped posture, a spiral posture and a wave-shaped posture; and the first posture is different from the second posture.

[0035] According to the specific embodiments of the application, the following technical effects are achieved:

[0036] The pipeline robot navigation method provided by the application, through the cooperative control of the artificial potential field method and the multi-joint unit pose adjustment, significantly improves the obstacle avoidance capability of the robot in the complex pipeline environment. The driving direction calculation based on the artificial potential field method can dynamically perceive the local obstacle distribution, combined with the real-time adjustment of the target angle, so that the robot can flexibly change the body posture (such as rotating around the pipeline axis or avoiding obstacles by turning sideways), effectively solving the problem of improper handling of repulsive force caused by the mechanical configuration limitation of the traditional method. This autonomous posture adjustment mechanism not only enables the robot to adapt to the complex environment of narrow pipeline sections and obstacle dense areas, but also enables the robot to quickly find a feasible path in a strong repulsive force field, avoiding stagnation caused by local obstacles. At the same time, through the cooperative movement of the multi-joint unit, the robot can take into account the obstacle shape and pipeline geometric characteristics during obstacle avoidance, significantly improving the avoidance efficiency of dynamic obstacles (such as flowing sediments), and ultimately enhancing the autonomy of the robot in the closed pipeline and the reliability of the detection and maintenance task. BRIEF DESCRIPTION OF DRAWINGS

[0037] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the figures, and wherein:

[0038] Figure 1 A method flowchart of the pipeline robot navigation method shown in the embodiment of the present application;

[0039] Figure 2 A schematic diagram of the artificial potential field distribution of the pipeline robot in the pipeline shown in the embodiment of the present application;

[0040] Figure 3 A local obstacle avoidance schematic diagram of the pipeline robot shown in the embodiment of the present application;

[0041] Figure 4 A schematic diagram of the loss value calculation principle when planning a global path shown in the embodiment of the present application;

[0042] Figure 5 A schematic diagram of the principle from global path planning to local path planning shown in the embodiment of the present application;

[0043] Figure 6 A schematic diagram of the module structure of the pipeline robot navigation system shown in the embodiment of the present application. DETAILED DESCRIPTION

[0044] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of the present application.

[0045] The purpose of the present application is to provide a pipeline robot navigation method and system, which can make the pipeline robot efficiently avoid obstacles in a complex pipeline environment through the artificial potential field method and multi-joint unit cooperative posture adjustment, and improve the autonomy and detection and maintenance efficiency of the robot.

[0046] Please refer to Figure 1 , Figure 1 The method flowchart of the pipeline robot navigation method.

[0047] The above-mentioned pipeline robot navigation method is applied to a processor of a pipeline robot, the pipeline robot includes a plurality of active joint units connected, and the navigation method includes:

[0048] Step 101: in response to the obstacle avoidance instruction, the target point attraction and the obstacle repulsion are calculated by using the artificial potential field method according to the target point position and the obstacle information.

[0049] In the application, when the obstacle avoidance instruction is received, the target point position and the obstacle information in the current environment need to be determined first. The obstacle avoidance instruction is triggered by the external sensor module and the internal sensor module. The external sensor identifies static obstacles and dynamic obstacles by collecting pipeline environment data (such as obstacle distance, shape and motion state). The internal sensor monitors the posture, speed and joint state of the robot body in real time. Based on the sensor fusion technology, the environment data and the robot state data are integrated and processed to generate obstacle information including obstacle position, size and motion trend, and the target point coordinates provided by the global path planning unit in the processor are combined to provide input for subsequent obstacle avoidance calculation.

[0050] Please refer to Figure 2 , Figure 2 In the figure, the labels 1, 2, 3, 4 and 5 represent the starting point, the target point, the obstacle, the repulsion field and the attraction field in sequence; the artificial potential field distribution of the pipeline robot during walking in the pipeline is shown in Figure 2 When the pipeline robot approaches the obstacle, the repulsion force given by the obstacle will be perceived, and the pipeline robot adjusts the body posture to avoid the obstacle and continues to move towards the target point.

[0051] In the framework of artificial potential field method, the target point is regarded as the source of attractive force, and the obstacle generates repulsive force. The attractive force is used to guide the robot to move towards the target point, and its size is related to the distance between the robot and the target point. When the robot is close to the target point, the attractive force increases linearly with the distance; when the distance exceeds the preset threshold, in order to avoid the problem that the attractive force is too large, the attractive force is limited by using an exponential decay function, so as to ensure that the robot moves more smoothly when approaching the target. The calculation of the repulsive force is dynamically adjusted according to the distance between the robot and the obstacle, and when the obstacle enters the preset repulsive force action range, the repulsive force rapidly increases, forcing the robot to adjust the path; when it exceeds the range, the repulsive force is zero to avoid unnecessary path interference. For dynamic obstacles, the future trajectory is predicted through a motion prediction model, and the direction and strength of the repulsive force are dynamically corrected, so that the robot can avoid moving obstacles in advance.

[0052] The attractive force and the repulsive force calculated by the artificial potential field method are combined in the vector space to form the total potential field, which directly determines the driving direction of the robot. In this process, the geometric characteristics of the pipeline and the mechanical configuration of the robot are considered comprehensively, for example, in narrow pipeline sections or obstacle-dense areas, the weight of the repulsive force will be adaptively increased to ensure that the robot preferentially selects a low-risk path. At the same time, the pipeline robot can flexibly adjust the body posture through the coordinated movement of the multi-joint unit, such as rotating around the pipeline axis or turning to avoid obstacles, to adapt to the avoidance needs of different shaped obstacles. This calculation process not only relies on real-time environmental perception data, but also synchronizes information with the global path planning unit to ensure that the local obstacle avoidance decision and the global path optimization are consistent, thereby improving the efficiency and reliability of the robot in complex pipeline environments.

[0053] wherein the calculation formula of the attractive force is a piecewise function, when d goal <d0, F att =k att ×d goal ; in the formula, F att is the attractive force, k att is the attractive force coefficient, d goal is the distance from the pipeline robot to the target point, and d0 is the attractive force distance threshold. When d goal ≥d0, in order to avoid the problem that the attractive force is too large when the distance from the target is large, the attractive force value is limited, and the attractive force The relationship between the attractive force field function U att and F att is Similarly, the relationship between the repulsive force field function U rep and F rep is that is, the force size is the negative gradient of the potential field, and the total potential field function U all =Uatt +U rep The repulsive force formula is also a piecewise function, when d obs < d1, When d obs > d1, F rep = 0, in the formula, F rep is the repulsive force, k rep is the repulsive force coefficient, d obs is the distance to the obstacle, d1 is the repulsive force distance threshold, is the unit vector from the obstacle to the robot. In the application, the attractive force and repulsive force field are calculated by the formula to provide the basis for local path planning. The robot in the pipeline is subjected to the resultant force of the external potential field in the artificial potential field F all = F att + F rep .

[0054] Step 102: Determine the driving direction and the target included angle of the adjacent joint unit based on the attractive force and the repulsive force.

[0055] In the application, after the attractive force of the target point and the repulsive force of the obstacle are calculated, the driving direction of the pipeline robot can be determined by vector superposition. The driving direction is determined by the direction of the total potential field, which comprehensively considers the guidance of the target point and the avoidance demand of the obstacle. For dynamic obstacles, the motion prediction model can be combined to correct the direction of the resultant force in real time, so as to ensure that the driving direction can predict and adapt to the moving trend of the obstacle. In this process, the mechanical configuration characteristics of the pipeline robot are taken into account, for example, in the narrow pipeline section or the obstacle dense area, the weight distribution of the driving direction is adaptively adjusted according to the inner diameter of the pipeline and the obstacle distribution, and the path that can minimize the friction or collision risk of the body is preferentially selected.

[0056] The target included angle of the adjacent joint unit is determined by the driving direction and the geometric characteristics of the pipeline. According to the type of the pipeline (such as straight pipe, elbow pipe, variable diameter pipe, etc.), the constraint conditions of different pipeline sections on the robot posture are predicted, and the expected motion trajectory of each joint unit is calculated combined with the current driving direction. For example, in the elbow section, the joint unit needs to form an arc-shaped posture to adapt to the curvature of the pipeline; in the straight pipe section, the joint unit maintains a straight-line posture to improve the marching efficiency. When encountering obstacles, the system dynamically plans the target included angle of the adjacent joint unit based on the repulsive force distribution and the shape of the obstacle. For example, for lateral obstacles, the robot is adjusted to rotate around the pipeline axis to form a spiral posture to avoid laterally; for front obstacles, a wave-shaped posture is formed by the coordinated bending of multiple joints to bypass the obstacle and then restore the original posture. In this process, the coordinated motion of each joint unit needs to ensure the overall stability of the body to avoid motion jamming or energy loss caused by local excessive bending.

[0057] The calculation of the target included angle can also involve a dynamic feedback mechanism. When the robot performs an obstacle avoidance action, the internal sensor module monitors the actual angles and motion states of each joint unit in real time and compares them with the target included angle. If the deviation exceeds a preset threshold, the travel direction and obstacle information are re-evaluated, and the target included angle parameters are dynamically adjusted to ensure the accuracy and reliability of the obstacle avoidance action. For example, in the scenario of a dynamic obstacle approaching quickly, the target included angle is iteratively optimized to shorten the posture adjustment time of the joint units and improve the response speed of obstacle avoidance.

[0058] Step 103: Adjusting the pipeline robot from the first posture to the second posture to avoid the obstacle according to the travel direction and the target included angle.

[0059] The first posture or the second posture is at least one of a straight posture, an S-shaped posture, an arc-shaped posture, a spiral posture, and a wave-shaped posture; and the first posture is different from the second posture.

[0060] In the application, after determining the travel direction and the target included angle of the adjacent joint units, the pipeline robot is dynamically adjusted from an initial posture (first posture) to an adaptive obstacle avoidance posture (second posture) according to the current pipeline environment characteristics and the obstacle distribution. The initial posture is determined by the motion state of the robot before entering the current pipeline section, for example, a straight posture is usually maintained in a straight pipeline section to reduce the resistance of travel; when an obstacle is detected, the posture adjustment mechanism is triggered based on the calculation results of the travel direction and the target included angle, and the coordinated motion of each joint unit is controlled to convert the robot from the initial posture to a new posture that matches the characteristics of the obstacle and the structure of the pipeline. For example, when a static lateral obstacle is encountered in a narrow elbow section, the robot forms a spiral posture by rotating around the pipeline axis to bypass the obstacle using a spiral travel path; if the obstacle is located directly in front and has an irregular shape, a wave-shaped posture is formed by alternating bending of multiple joints to achieve continuous avoidance.

[0061] During the posture adjustment process, the motion parameters of each joint unit are accurately controlled by the target included angle. The target included angle is dynamically generated according to the travel direction and the relative position of the obstacle to ensure that the bending angles of adjacent joint units adapt to the curvature of the pipeline and the profile of the obstacle. For example, when the robot needs to pass through a curved pipeline section in an arc-shaped posture, the target included angle is set to make the joint units form a smooth arc arrangement along the inner wall of the pipeline to reduce friction between the body and the pipe wall; when facing a dynamic obstacle, the target included angle is updated in real time in combination with a motion prediction model to make the robot quickly adjust the travel trajectory through an S-shaped posture to avoid the interference of the moving obstacle. The execution of posture conversion needs to consider both motion efficiency and stability to avoid body shaking or energy loss caused by excessive joint action amplitude.

[0062] Referring to Figure 3 , Figure 3 Reference numerals 3 and 6 represent the obstacle and the rotation axis, respectively; asFigure 3 As shown, the pipe robot can rotate around the pipe axis by rotating the front and rear spherical wheels in the same direction, and adjust the posture of the pipe robot to a spiral posture; the pipe robot can also adjust the posture by turning sideways, such as adjusting to an arc-shaped posture, to avoid obstacles during the process of passing through static obstacles inside the pipe.

[0063] The new adjusted posture needs to meet different shape requirements from the initial posture to ensure the effectiveness of the obstacle avoidance action. For example, if the initial posture is a straight-line posture, it can be converted to a spiral posture or a wave-shaped posture during obstacle avoidance; if the initial posture is an arc-shaped posture, it can be switched to an S-shaped posture to adapt to the spatial distribution of sudden obstacles. The core of the posture differentiation is to change the overall configuration of the robot through the coordinated deformation of the joint units, so as to open up a new feasible path in the limited pipe space. In this process, the angle change and stress state of each joint unit can be monitored in real time, so as to perform dynamic correction and ensure the accuracy and controllability of the posture adjustment. In addition, the updated posture information can be synchronized to the global path planning unit for optimizing the motion strategy of the subsequent path segment, forming a cooperative closed loop of global and local planning, and finally realizing efficient obstacle avoidance and stable passage of the robot in a complex pipe environment.

[0064] In one embodiment, the target point position is from a remote end, and the above navigation method further comprises:

[0065] Based on the loss value of the predefined pipe structure feature and the target point position, the total loss function of each candidate path in the pipe global map is calculated by a hybrid A-star algorithm, and the global path is determined by minimizing the total loss function.

[0066] In the application, the target point position can be sent by an operator through a remote end (such as a cloud, an upper computer), or can be issued to the robot system by the remote end based on task requirements, such as detecting the end point of the task or specifying the area of the maintenance work. The remote end transmits the coordinate information of the target point and the task priority parameter to the global path planning unit of the robot through the communication module. In order to generate a global path, loss values need to be predefined for different types of pipe segments according to pipe structure features, for example, the straight pipe segment has a higher passage efficiency, and the loss value is set to a lower weight; the variable-diameter pipe or tee pipe has a complex structure and large passage resistance, and the loss value is correspondingly increased. These loss values reflect the comprehensive influence of different pipe types on the motion efficiency, energy consumption and passage probability of the robot, and provide a basis for quantitative evaluation of the global path.

[0067] The application of the hybrid A-star algorithm in the global path planning contains a multi-stage calculation process. Based on the position of the target point, one or more candidate paths are selected from the pre-stored global map of the pipeline. When there are multiple candidate paths, the total loss function of each candidate path is obtained by weighted summation of the loss values of each pipeline section, and is dynamically adjusted in combination with the time cost and the probability of passability. For example, in a narrow elbow section, the algorithm increases the weight of the time cost to avoid path congestion; in an obstacle-dense area, the weight of the probability of passability is increased to ensure the feasibility of the path. Through iterative optimization, the path with the minimum total loss function is selected as the global planning result. This path not only covers the general framework of the direction of the target point, but also marks the characteristic parameters (such as the radius of curvature and the change of the inner diameter) of the key pipeline sections, providing prior information for the subsequent local path planning.

[0068] The global path provides macro guidance for local obstacle avoidance decision, such as adjusting the robot pose in advance when approaching a variable-diameter pipe section; the changes in the distribution of obstacles or newly opened efficient path sections found in the local planning process can be fed back to the global path planning unit in real time, triggering dynamic updating and path re-planning of the global map. This closed-loop mechanism ensures that the robot can not only follow the globally optimized path, but also flexibly respond to local unexpected situations, ultimately achieving reliable navigation from the starting point to the target point.

[0069] Referring to Figure 4 , Figure 4 FIG. 1 is a schematic diagram of the principle of loss value calculation for global path planning.

[0070] In the application, the hybrid A-star algorithm is used for global path planning of the pipeline robot. The loss values of different pipeline types are defined in advance, for example, the loss value of each meter of straight pipeline cost(straight pipe) = 1, the loss value of each variable-diameter pipeline cost(variable diameter) = 2, the loss value of each three-way pipeline cost(three-way pipe) = 5, and the loss value of each elbow pipeline cost(elbow pipe) = 3. By dividing the pipeline types and setting the loss values, the total loss function value ∑cost is calculated during the execution of the hybrid A-star global path planning algorithm, that is, the loss values of each pipeline type are weighted and summed, the objective function is based on the shortest time and the maximum probability of passability, and the optimal global path planning route is selected based on the calculation results of the loss functions of different paths; the present application takes into account the influence of different pipeline types on the robot's progress, making the path planning more suitable for the actual pipeline environment.

[0071] Referring to Figure 5 , Figure 5 FIG. 2 is a schematic diagram of the principle from global path planning to local path planning; Figure 5 In FIG. 1, reference numerals 1, 2 and 3 represent the starting point, the target point and the obstacle, respectively.

[0072] In the application, the global planning is used to roughly determine the route, and the improved artificial potential field method is used for local obstacle avoidance planning to ensure the smooth movement of the robot. For example Figure 5 As shown in the figure, the global path planning route is obtained by hybrid A-star, and then local motion planning is further performed. The green line on the left side of the figure determines the rough movement direction of the robot from the starting point to the end point. Then, the improved artificial potential field method is used for local path planning. The middle part of the figure introduces the method used for local path planning, and gives the calculation formula of the attractive force and the repulsive force. Finally, the obstacle avoidance motion is performed according to the force field calculated in the foregoing, so that the pipe robot avoids the obstacles and improves the performance of the movement in the pipe.

[0073] In one embodiment, the above navigation method further comprises:

[0074] When the pipe robot successfully avoids the obstacles, the obstacle distribution data and the effective path segment are generated according to the corresponding obstacle information, the movement direction and the target angle.

[0075] Based on the obstacle distribution data and the effective path segment, the global map of the pipe is updated.

[0076] In the application, after the pipe robot successfully completes the obstacle avoidance, the obstacle distribution data can be generated based on the obstacle position, size and motion state information collected during the obstacle avoidance, combined with the movement direction of the robot and the target angle parameters of the joint unit. The data includes the spatial coordinates of the obstacle, the type identifier (such as static obstacle or dynamic obstacle) and the motion trend prediction result, for example, the velocity vector and potential trajectory of the dynamic obstacle. At the same time, the effective path segment information of the actual passage during the obstacle avoidance stage is extracted, and the geometric characteristics (such as curvature and internal diameter change) and joint motion parameters (such as the target angle sequence of each joint unit) of the path segment are recorded, which provides reusable local passage experience for subsequent path planning.

[0077] The generated obstacle distribution data and effective path segment can be fed back to the global path planning unit to trigger the real-time update of the global map of the pipe. During the update process, the global map will add or correct the coordinate range and attribute label of the obstacle distribution area, for example, mark the active period of the dynamic obstacle or the safe bypass distance of the static obstacle. For the effective path segment, it can be compared and analyzed with the original candidate path library. If the passage efficiency or safety of the path segment is better than that of the historical path, it will be included in the preferred set of the global path as an important reference for subsequent planning. For example, the joint angle sequence and movement direction parameters of the spiral path successfully avoiding the obstacle in the narrow curved pipe segment will be stored and used for rapid path generation in similar pipe environments.

[0078] The updating mechanism of the global map can adopt an incremental data fusion strategy to ensure the compatibility of new information and historical data. When multiple valid path segments overlap, the actual travel time, energy consumption, and obstacle avoidance success rate of the path segments are weighted and evaluated to dynamically adjust the path priority. At the same time, the updating of obstacle distribution data and path segment information triggers the adaptive optimization of the parameters of the hybrid A-star algorithm, such as increasing the penalty weight of the loss function in obstacle-dense areas or reducing the time cost coefficient in high-passability path segments. This closed-loop feedback mechanism enables the global path planning to continuously iterate based on actual obstacle avoidance experience, improving the navigation efficiency and robustness of the robot in complex pipeline environments, and ultimately realizing the collaborative evolution from local obstacle avoidance to global optimization.

[0079] In one embodiment, the pipeline robot further comprises an internal sensor module and an external sensor module; and the method for generating the obstacle avoidance instruction comprises:

[0080] acquiring state data of the pipeline robot body collected by the internal sensor module and pipeline environment data collected by the external sensor module;

[0081] determining obstacle information based on the pipeline environment data;

[0082] generating the obstacle avoidance instruction according to the state data and the obstacle information.

[0083] In the application, the generation of the obstacle avoidance instruction can be based on the cooperative work of the internal sensor module and the external sensor module. The internal sensor module is responsible for collecting real-time state data of the robot body, including the motion angle of the joint unit, the body attitude parameters (such as the pitch angle and the yaw angle), the travel speed, and the torque feedback of each joint. These data reflect the current motion state of the robot and the stability of the mechanical configuration, such as whether the body is abnormally tilted in a narrow pipeline segment or whether the bending angle of the joint unit exceeds the safety threshold. The external sensor module collects pipeline environment data through a laser radar, a vision camera, or an ultrasonic sensor, including the distance, the contour size, the surface features, and the motion state (such as the fixed coordinates of static obstacles or the velocity vector of dynamic obstacles) of obstacles.

[0084] Based on the pipeline environment data obtained by the external sensor module, the spatial distribution characteristics of the obstacles are identified through a multi-source data fusion algorithm. For example, laser radar point cloud data is used to construct a three-dimensional profile of the obstacle, and a vision sensor distinguishes the type of the obstacle (such as a weld protrusion or a sediment accumulation) through image recognition technology, and an ultrasonic sensor assists in measuring the relative motion speed of a dynamic obstacle. In combination with time series analysis, the attributes (static or dynamic) of the obstacle and its potential threat level are further determined, such as whether the motion trajectory of the dynamic obstacle intersects with the current path of the robot. The finally generated obstacle information can include the coordinates, type, size, and motion trend prediction results of the obstacle, providing a quantitative basis for subsequent obstacle avoidance decisions.

[0085] The generation of the obstacle avoidance instruction can dynamically associate the state data of the internal sensor module with the obstacle information. For example, when the robot is in a high-speed travel state, if a close-range obstacle is detected in front, the feasible obstacle avoidance action response time and attitude adjustment amplitude can be calculated in combination with the current speed, the adjustable angle range of the joint, and the inner diameter size of the pipeline. If the internal sensor feedbacks that the body attitude is close to the stability limit (such as a large inclination angle), a low-amplitude and gradual attitude adjustment strategy can be preferentially selected to avoid motion out of control due to excessive action.

[0086] In one embodiment, the obstacle information indicates that the obstacle is a dynamic obstacle, and the artificial potential field method is used to calculate the obstacle repulsion force, including:

[0087] Based on the obstacle information, a pre-trained motion prediction model is used to generate the motion trajectory of the dynamic obstacle.

[0088] According to the motion trajectory, the repulsion force calculation parameters in the artificial potential field method are adjusted to dynamically calculate the obstacle repulsion force.

[0089] In application, when the obstacle information indicates that the obstacle is a dynamic obstacle, the repulsion force calculation strategy can be dynamically adjusted through the cooperative mechanism of the motion prediction model and the artificial potential field method. A dynamic obstacle refers to an object that has the ability to move within the pipeline, such as flowing sediment or a moving detection device, whose position changes over time and can continuously interfere with the robot's path. For such obstacles, the speed, acceleration, and direction information of the dynamic obstacle can be determined based on the real-time pipeline environment data collected by the external sensor module, and a pre-trained motion prediction model is called to generate the motion trajectory prediction results of the dynamic obstacle in the future time period through time series data analysis and trajectory fitting algorithms. This motion prediction model can be trained through historical pipeline environment data and reinforcement learning, and can output a high-confidence trajectory prediction curve by combining the obstacle type, pipeline geometric constraints, and fluid mechanics characteristics, such as predicting the displacement path and possible dangerous area covered by the dynamic obstacle in the next 5 seconds.

[0090] The calculation parameter of the repulsive force in the artificial potential field method can be dynamically corrected according to the predicted trajectory of the dynamic obstacle. The traditional calculation of the repulsive force only considers the current obstacle position, and for the dynamic obstacle, the present application can discretize the predicted trajectory into multiple time node potential threat points, and assign a dynamic weight coefficient to each threat point. For example, the closer the threat point is to the current time, the higher the weight coefficient, and the corresponding repulsive force range extends to the area covered by the predicted trajectory. At the same time, the direction of the repulsive force is no longer only directed to the current obstacle position, but is distributed along the normal direction of the predicted trajectory, forming a fan-shaped repulsive force field, so that the robot can avoid the moving path of the dynamic obstacle in advance. In addition, the repulsive force coefficient can be adaptively adjusted according to the obstacle movement speed: if the dynamic obstacle approaches the robot quickly, the repulsive force coefficient is increased to enhance the avoidance response; if the obstacle is in a low-speed or away state, the coefficient is reduced to avoid excessive path deviation.

[0091] Through the above dynamic adjustment mechanism, the repulsive force generated by the artificial potential field method can effectively deal with the uncertainty of the dynamic obstacle. For example, when the prediction model judges that the dynamic obstacle will move to the left along the pipeline axis, the strength of the repulsive force field in the left area is significantly improved, guiding the robot to adjust the driving direction to the right side or change the body posture (such as forming a spiral posture to bypass). At the same time, the system combines the real-time motion state of the pipeline robot (such as speed, joint adjustable angle) and the time window of the predicted trajectory to dynamically optimize the effective time length of the repulsive force action, ensuring that the obstacle avoidance action is synchronized with the obstacle movement rhythm. This strategy enables the robot to avoid immediate collision risks in complex dynamic environments while reducing path detours caused by excessive avoidance, ultimately achieving efficient and safe autonomous navigation.

[0092] In one embodiment, the first posture is specifically a straight posture, and the second posture is specifically a spiral posture, and the pipeline robot is adjusted from the first posture to the second posture to avoid the obstacle according to the driving direction and the target angle, including:

[0093] According to the driving direction and the target angle, the joint units are controlled to rotate around the pipeline axis direction in cooperation, so that the pipeline robot is converted from the straight posture to the spiral posture, forming a spiral-shaped travel path to bypass the obstacle.

[0094] In the application, when the first posture is a straight posture and needs to be converted to a second posture (a spiral posture) to avoid the obstacle, the cooperative motion control mechanism of the multi-joint unit can be triggered based on the calculation results of the driving direction and the target angle. In the straight posture, the joint units of the robot are linearly arranged along the pipeline axis, the travel path is parallel to the axis, and it is suitable for efficient passing through straight pipe sections; and in the spiral posture, the joint units are required to rotate around the pipeline axis direction in synchronization, forming a spiral-shaped body configuration, so that the robot travels along a spiral trajectory, thereby bypassing the lateral or front obstacle.

[0095] The execution of the posture conversion is based on the dynamic allocation of the target angle; according to the relative position of the obstacle and the geometric characteristics of the pipeline, the target angle sequence of the rotation of each joint unit around the axis is calculated in real time. For example, if the obstacle is located on the right side of the pipeline, the joint unit needs to rotate in the clockwise direction in turn, so that the whole body of the robot presents a right-handed helical configuration, and the robot is pushed to the left front to avoid the obstacle. During the rotation process, the angular velocity and phase difference of each joint unit are strictly matched to ensure the continuity and stability of the helical posture, and to avoid the distortion or motion jamming of the body due to local rotation lag.

[0096] The generation of the helical travel path depends on the coordinated cooperation of the rotation of the joint unit and the axial travel. After the linear posture conversion into the helical posture, the robot maintains the axial propulsion force, and at the same time forms a helical trajectory through the continuous rotation of the joint unit. The pitch and radius of the trajectory can be dynamically adjusted according to the target angle and the inner diameter of the pipeline, for example, the helical radius is appropriately reduced in the narrow pipeline section, and the pitch is increased in the obstacle-dense area to improve the obstacle avoidance efficiency. The advantage of the helical posture in obstacle avoidance is that the travel path can bypass the obstacle while maintaining uniform contact with the pipeline wall, reducing local friction resistance, and enhancing the stability of the body by utilizing the centrifugal force effect. After the posture adjustment is completed, the maintenance state of the helical configuration can be monitored in real time through the internal sensor, and if the posture deviation or external interference is detected, the target angle is recalculated and the joint rotation parameters are fine-tuned to ensure the accurate execution of the obstacle avoidance action and the reliability of the path.

[0097] Corresponding to the foregoing application function implementation method embodiment, the present application also provides a pipeline robot navigation system and corresponding embodiments.

[0098] Please refer to Figure 6 , Figure 6 for a schematic diagram of the module structure of the pipeline robot navigation system.

[0099] The above-mentioned pipeline robot navigation system is applied to the processor of the pipeline robot, the pipeline robot comprises a plurality of actively connected joint units, and the system comprises:

[0100] The obstacle avoidance decision unit 61 is configured to, in response to the obstacle avoidance instruction, calculate the target point attractive force and the obstacle repulsive force by using the artificial potential field method according to the target point position and the obstacle information;

[0101] The motion planning unit 62 is configured to determine the travel direction and the target angle of the adjacent joint unit based on the attractive force and the repulsive force;

[0102] The posture control unit 63 is configured to adjust the pipeline robot from a first posture to a second posture according to the travel direction and the target angle to avoid the obstacle; the first posture or the second posture is at least one of a linear posture, an S-shaped posture, an arc-shaped posture, a helical posture and a wave-shaped posture; the first posture is different from the second posture.

[0103] In one embodiment, the target point position is from a remote end, and the navigation system further comprises:

[0104] a global path planning unit, configured to calculate a total loss function of each candidate path in the global pipe map based on the loss value of the predefined pipe structure feature and the target point position, and determine a global path by minimizing the total loss function.

[0105] In one embodiment, the navigation system further comprises:

[0106] an updating unit, configured to:

[0107] generate obstacle distribution data and an effective path segment according to the corresponding obstacle information, the driving direction and the target angle when the pipe robot successfully avoids the obstacle;

[0108] update the global pipe map based on the obstacle distribution data and the effective path segment.

[0109] In one embodiment, the pipe robot further comprises an internal sensor module and an external sensor module, and the navigation system further comprises:

[0110] an instruction generation unit, configured to:

[0111] obtain state data of the pipe robot body collected by the internal sensor module and pipe environment data collected by the external sensor module;

[0112] determine obstacle information based on the pipe environment data;

[0113] generate an obstacle avoidance instruction according to the state data and the obstacle information.

[0114] In one embodiment, the obstacle information indicates that the obstacle is a dynamic obstacle, and the obstacle avoidance decision unit 61 is specifically configured to:

[0115] generate a motion trajectory of the dynamic obstacle by using a pre-trained motion prediction model based on the obstacle information;

[0116] adjust a repulsive force calculation parameter in the artificial potential field method according to the motion trajectory, so as to dynamically calculate the repulsive force of the obstacle.

[0117] In one embodiment, the first posture is a straight-line posture, and the second posture is a spiral posture, and the posture control unit 63 is specifically configured to:

[0118] According to the driving direction and the target included angle, control each joint unit to rotate collaboratively around the pipeline axis direction, so that the pipeline robot is converted from a straight-line posture to a spiral posture, forming a spiral traveling path to bypass obstacles.

[0119] In one embodiment, the above obstacle avoidance decision unit 61 uses the following calculation formula to calculate the attraction force and the repulsion force:

[0120] The calculation formula for the attraction force is:

[0121] When d goal < d0, F att = k att × d goal ;

[0122] When d goal ≥ d0,

[0123] In the formula, F att is the attraction force, k att is the attraction coefficient, d goal is the distance from the pipeline robot to the target point, and d0 is the attraction distance threshold;

[0124] The calculation formula for the repulsion force is:

[0125] When d obs < d1,

[0126] When d obs > d1, F rep = 0;

[0127] In the formula, F rep is the repulsion force, k rep is the repulsion coefficient, d obs is the distance from the pipeline robot to the obstacle, d1 is the repulsion distance threshold, is the unit vector pointing from the obstacle to the robot.

[0128] Regarding the system in the above embodiment, the specific manner in which each unit performs operations has been described in detail in the embodiment related to the method, and will not be elaborated here.

[0129] Having described various embodiments of the application, it is to be understood that the above description is meant to be illustrative only and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art, without departing from the scope and spirit of the described embodiments. The choice of words in this document is intended to best explain the principles of the embodiments, the practical application, or improvement over the technology in the art, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method of navigating a pipe robot, characterized by, The pipeline robot comprises a plurality of articulated joint units, and the method comprises: in response to the obstacle avoidance instruction, calculating target point attraction and obstacle repulsion by using an artificial potential field method according to the target point position and the obstacle information; based on the attraction and the repulsion, determining a driving direction and a target included angle of adjacent joint units; the target included angle of adjacent joint units is determined by the driving direction and the geometric characteristics of the pipeline; according to the type of the pipeline, the constraint conditions of different pipeline sections on the posture of the robot are predicted, and the expected motion trajectory of each joint unit is calculated in combination with the current driving direction; according to the driving direction and the target included angle, adjusting the pipeline robot from a first posture to a second posture to avoid the obstacle; the first posture or the second posture is at least one of a straight-line posture, an S-shaped posture, an arc-shaped posture, a spiral posture and a wave-shaped posture; the first posture is different from the second posture; in the posture adjustment process, the motion parameters of each joint unit are controlled by the target included angle; the calculation formula of the attraction is: When Time, ; When Time, ; wherein, is an attractive force, is an attractive force coefficient, is a distance of the pipe robot to the target point, is an attractive force distance threshold; the calculation formula of the repulsion is: When Time, ; When Time, ; wherein is the repulsive force, is the repulsive force coefficient, is the distance of the pipe robot to the obstacle, is the repulsive force distance threshold, is the unit vector pointing from the obstacle to the robot.

2. The method of claim 1, wherein, the target point position is from a remote end, and the method further comprises: based on the loss value of the pre-defined pipeline structure characteristics and the target point position, calculating a total loss function of each candidate path in the pipeline global map by using a hybrid A-star algorithm, and determining a global path with the minimum total loss function as the target.

3. The method of claim 1, wherein, the method further comprises: when the pipeline robot successfully avoids the obstacle, generating obstacle distribution data and an effective path section according to the corresponding obstacle information, the driving direction and the target included angle; based on the obstacle distribution data and the effective path section, updating the pipeline global map.

4. The method of claim 1, wherein, The pipeline robot further comprises an internal sensor module and an external sensor module; the method for generating the obstacle avoidance instruction comprises: acquiring state data of the pipeline robot body collected by the internal sensor module and pipeline environment data collected by the external sensor module; based on the pipeline environment data, determining obstacle information; generating an obstacle avoidance instruction according to the state data and the obstacle information.

5. The method of Claim 1, wherein, If the obstacle information indicates that the obstacle is a dynamic obstacle, the repulsion of the obstacle is calculated by using the artificial potential field method, which comprises: based on the obstacle information, generating the motion trajectory of the dynamic obstacle by using a pre-trained motion prediction model; adjusting the repulsion calculation parameters in the artificial potential field method according to the motion trajectory, so as to dynamically calculate the repulsion of the obstacle.

6. The method of Claim 1, wherein, If the first posture is a straight-line posture and the second posture is a spiral posture, adjusting the pipeline robot from the first posture to the second posture to avoid the obstacle according to the driving direction and the target included angle, which comprises: controlling each joint unit to rotate cooperatively around the axis direction of the pipeline according to the driving direction and the target included angle, so that the pipeline robot is converted from the straight-line posture to the spiral posture, and a spiral-shaped advancing path is formed to bypass the obstacle.

7. A pipeline robot navigation system characterized by, The pipeline robot comprises a plurality of articulated joint units, and the system comprises: an obstacle avoidance decision unit, configured to calculate target point attraction and obstacle repulsion by using an artificial potential field method according to the target point position and the obstacle information in response to the obstacle avoidance instruction; a motion planning unit, configured to determine a driving direction and a target included angle of adjacent joint units based on the attractive force and the repulsive force, and to predict constraint conditions of different pipe sections on the robot pose according to the pipe type, and to calculate an expected motion trajectory of each joint unit according to the driving direction and the constraint conditions; a pose control unit, configured to adjust the pipe robot from a first pose to a second pose according to the driving direction and the target included angle, so as to avoid the obstacle; the first pose or the second pose is at least one of a straight pose, an S-shaped pose, an arc-shaped pose, a spiral pose and a wave-shaped pose; the first pose is different from the second pose; motion parameters of each joint unit in the pose adjustment process are controlled by the target included angle; a calculation formula of the attractive force is: When Time, ; When Time, ; wherein, is an attractive force, is an attractive force coefficient, is a distance of the pipe robot to the target point, is an attractive force distance threshold; a calculation formula of the repulsive force is: When time, ; When Time, ; wherein is the repulsive force, is the repulsive force coefficient, is the distance of the pipe robot to the obstacle, is the repulsive force distance threshold, is the unit vector pointing from the obstacle to the robot.

Images