Tower assembling robot control method and system, tower assembling robot and tower assembling system

Abstract

The invention provides a tower assembling robot control method and system, a tower assembling robot and a tower assembling system, and relates to the technical field of electric power construction, and the tower assembling robot control method comprises the following steps: respectively obtaining images collected under a first focal length and a second focal length; fusing the images collected under the first focal length and the second focal length to obtain a fused image, extracting geometric features of the bolt based on the fused image, and estimating an initial pose of the bolt based on the extracted geometric features of the bolt; optimizing the initial pose in combination with a pose optimization strategy to obtain an optimized bolt pose; and the tower assembling robot is controlled to conduct fastening operation on the to-be-fastened bolt by combining the optimized bolt pose and the self-adaptive impedance control bolt fastening scheme. The tower assembling efficiency of the high-voltage power transmission iron tower can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power construction, and in particular to a tower assembly robot control method and system, a tower assembly robot, and a tower assembly system. Background Art

[0002] During the on-site assembly and construction of high-voltage transmission towers, to avoid the high-altitude work risks associated with manual tower climbing, tower assembly robots can be used to replace manual tower climbing to perform bolt tightening and complete the tower assembly task. In related technologies, climbing tower assembly robots are often used to climb along the tower to the work position to tighten the bolts on the tower's angle steel. However, the tower itself is complex in structure, and the robot must climb over angle steel in different positions and orientations during climbing, which requires a high level of obstacle-crossing capability from the climbing mechanism. The entire climbing process also consumes a lot of power, limiting the robot's operating time on the tower and thus affecting construction progress. During the hoisting process, the tower assembly robot is hoisted to a set position on the high-voltage transmission tower and fixed there, then performs bolt tightening operations. However, since the relative position between the tower assembly robot and the transmission tower continues to fluctuate significantly, it is difficult for the tower assembly robot to accurately tighten the bolts on the transmission tower. Summary of the Invention

[0003] In order to address the deficiencies of the prior art, the purpose of the present invention is to provide a tower assembly robot control method, system, tower assembly robot and tower assembly system. The tower assembly robot can be moved to a set position on a high-voltage transmission tower by hoisting and fixed, and then bolt tightening operations are performed, thereby improving the tower assembly efficiency of the high-voltage transmission tower.

[0004] To achieve the above objectives, according to some embodiments, a first aspect of the present invention provides a tower assembly robot control method, comprising the following steps: Based on the tower assembly robot's three-point fixation and balance control method, the image and force sensor data at the preset position of the tower assembly robot are analyzed to generate a three-point fixation and balance strategy for the tower assembly robot and the tower, so as to achieve the fixation and balance of the tower assembly robot on the working surface and the tower. Based on the images obtained after the tower assembly robot and the tower are fixed, the geometric features of the corresponding bolts are extracted. Combined with the multi-type and multi-angle tower bolt initial positioning method based on the bolt geometric features, the coarse-grained working position of the robot is determined. Based on the coarse-grained working position of the robot and combined with the bolt and tool matching method of adaptive impedance control, the operations of multiple bolts and the corresponding fine-grained working positions of the robot are generated to control the tower assembly robot to tighten the bolts to be tightened.

[0005] Furthermore, based on the three-point fixation method of the tower assembly robot, the image at the preset position of the tower assembly robot is analyzed to generate a three-point fixation strategy for the tower assembly robot and the tower to achieve the fixation of the tower assembly robot on the working surface and the tower, including: identifying a target clamping area for a first image acquired at a first focal length; A focal length optimization objective function is constructed by maximizing the image average gradient and minimizing the image depth estimation error and the focal length variation error. After obtaining the optimal focal length, a second image of the target clamping area is acquired at the optimal focal length. The weight of the target image at the current optimal focal length is calculated, and the first image and the second image are weightedly fused. The target clamping area in the obtained fused image is used to identify the clamping point to control the clamping action of the gripper on the tower assembly robot on the tower body.

[0006] Furthermore, based on the tower assembly robot balance control method, the force sensor data at the preset position of the tower assembly robot is analyzed to generate a tower assembly robot and tower balance strategy to achieve the balance of the tower assembly robot after the working surface and the tower are fixed, including: The average force value detected by the first force sensor and the second force sensor is compared with the average force value detected by the third force sensor and the fourth force sensor, and the lifting ring is adjusted toward the side with the larger average force value by the lifting machine until the average forces at both ends are equal or the difference between the average forces at both ends is less than a preset difference; At the same time, the detection value of the first force sensor and the detection value of the second force sensor, as well as the detection value of the third force sensor and the detection value of the fourth force sensor are judged, and the lifting ring is adjusted to the side with the larger detection value through the lifting machine until the two compared detection values ​​are equal or the difference between the values ​​is less than the preset detection value difference.

[0007] Furthermore, the method of extracting corresponding bolt geometric features based on the images obtained after the tower assembly robot and the tower are fixed, and combining the multi-type and multi-angle tower bolt initial positioning method with the bolt geometric features to determine the coarse-grained working position of the robot includes: Acquire images captured at the first focal length and the second focal length after the tower assembly robot and the tower are fixed; The images captured at the first focal length and the second focal length are fused to obtain a fused image. The geometric features of the bolt are extracted based on the fused image, and the initial pose of the bolt is estimated based on the extracted geometric features of the bolt. The initial pose is optimized by combining the pose optimization strategy to obtain the optimized bolt pose. The optimized bolt pose is used to control the clamping action of the gripper on the tower assembly robot on the tower body.

[0008] Furthermore, when acquiring the image captured at the second focal length, a focal length optimization objective function is constructed by maximizing the image average gradient and minimizing the image depth estimation error and the focal length change error. After obtaining the optimal focal length, a second image of the target clamping area captured at the second focal length is acquired.

[0009] Furthermore, the focal length optimization objective function is: ; in, 、 and is the weight coefficient; At the current focal length The depth estimation value of the image under is the mean depth of historical images; is the average gradient of the image; and is the maximum focal length and the minimum focal length; is the center focal length.

[0010] Furthermore, the geometric features of the corresponding bolts are extracted based on the images obtained after the tower assembly robot and the tower are fixed, and the coarse-grained working position of the robot is determined by combining the multi-type and multi-angle tower bolt initial positioning method based on the geometric features of the bolts, including: extracting the geometric features of the bolts based on the acquired image data; The initial position of the bolt is estimated based on its geometric features. The coordinate position of the 3D point of the bolt is optimized by minimizing the reprojection error based on the BA optimization algorithm. The position of the clamp relative to the bolt to be tightened is determined based on the position of the bolt to be tightened and the position of the clamp. Then, the position of the bolt tightening device relative to the bolt to be tightened is obtained based on the position of the bolt tightening device relative to the clamp.

[0011] Furthermore, the bolt and tool matching method combined with adaptive impedance control generates multiple bolt operations and tool matching corresponding robot fine-grained operation positions, including: Construct the expected impedance equation; The stiffness of the environment is updated in real time to capture the stiffness changes during the bolt screwing phase. The expected impedance equation is updated based on the damping adjustment rule and the stiffness adjustment rule to obtain an updated expected impedance equation. The position is corrected based on the updated expected impedance equation and environmental stiffness to obtain the expected position, thereby controlling the tower assembly robot to tighten the bolts to be tightened.

[0012] Furthermore, the stiffness adjustment rule and the damping adjustment rule and include: , in, represents the initial stiffness, Represents the adaptive gain coefficient, when When the force is less than the expected value, increase To enhance the response, Greater than expected When Avoid overloading; , in, represents the initial damping, It represents the damping attenuation coefficient, which dynamically adjusts the damping according to the torque change rate to suppress oscillation.

[0013] A second aspect of the present invention provides a tower assembly robot control system, comprising: A fixation and balance control module is used to analyze the image and force sensor data at the preset position of the tower assembly robot based on the three-point fixation and balance control method of the tower assembly robot, and generate a three-point fixation and balance strategy for the tower assembly robot and the tower to achieve fixation and balance of the tower assembly robot on the working surface and the tower; The bolt pose optimization module is used to extract the corresponding bolt geometric features based on the images obtained after the tower assembly robot and the tower are fixed. It combines the multi-type and multi-angle tower bolt initial positioning method with the bolt geometric features to determine the coarse-grained working position of the robot; The tightening control module is used to generate multiple bolt operations and the corresponding fine-grained robot operation positions based on the robot's coarse-grained operation positions and the bolt and tool matching method with adaptive impedance control to control the tower assembly robot to tighten the bolts to be tightened.

[0014] To achieve the above objectives, according to some embodiments, a third aspect of the present invention provides a tower assembly robot, comprising: a robot body frame, a fixing mechanism and a bolt fastening mechanism, a zoom camera, and a control module; The robot body frame includes a frame and a lifting mechanism connected to the upper part of the frame; The fixing mechanism is arranged at the bottom of the frame and is used to fix the robot body frame to the tower body of the transmission tower; The bolt fastening mechanism includes a sliding shaft and a fastening tool, one end of the sliding shaft is slidably connected to the frame, and the other end is rotatably connected to the fastening tool, which is used to change the relative position of the fastening tool and the transmission tower, and the transmission tower is assembled by the fastening tool; The zoom camera is arranged at the fixing mechanism, and the control module is arranged on the robot body frame. The zoom camera is used to obtain images collected at a first focal length and a second focal length; The control module is configured to execute the tower assembly robot control method as described in the first embodiment.

[0015] Preferably, the sliding axis includes a longitudinal axis and a transverse axis, one end of the longitudinal axis is slidably connected to the frame, the transverse axis is slidably connected to an end of the longitudinal axis away from the frame, and the transverse axis is perpendicular to the longitudinal axis.

[0016] Preferably, the fastening tool is connected to the transverse axis via a sliding bracket, a first end of the sliding bracket is slidably connected to an end of the transverse axis away from the longitudinal axis, and a second end of the sliding bracket is rotatably connected to the fastening tool.

[0017] Preferably, there are at least two rotational degrees of freedom between the fastening tool and the sliding bracket, namely, horizontal rotation and vertical rotation.

[0018] Preferably, there are multiple fastening tools, and each fastening tool can rotate independently relative to the sliding bracket.

[0019] Preferably, the fixing mechanism includes a first fixing mechanism and a second fixing mechanism, and the first fixing mechanism and the second fixing mechanism are respectively arranged at two ends of the bottom of the frame and can slide relative to the frame.

[0020] Preferably, the frame includes a main beam and extensions arranged at both ends of the main beam, the extensions extending upward to a set height higher than the main beam; the top of the lifting mechanism has a lifting ring, and the two ends of the lifting mechanism are respectively connected to the tops of the extensions at both ends of the main beam.

[0021] A fourth aspect of the present invention provides a tower assembly system, comprising a hoisting device and the tower assembly robot provided in the first aspect, wherein the hoisting device has a lifting arm capable of cooperating with the hoisting mechanism of the tower assembly robot.

[0022] Preferably, there are multiple tower assembly robots, and the frames of the multiple tower assembly robots are fixedly connected in the length direction.

[0023] A fifth aspect of the present invention, based on the tower assembly robot provided in the first aspect or the tower assembly system provided in the second aspect, further provides a tower assembly method, comprising: Hoist the tower assembly robot to the set position of the transmission tower; Fix the tower assembly robot on the transmission tower through the fixing mechanism of the tower assembly robot; The sliding shaft moves, driving the fastening tool to the position of the bolt to be fastened on the transmission tower; The tightening tool moves to tighten the bolt to be tightened.

[0024] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention creatively proposes a tower assembly robot control method. First, the image and force sensor data are integrated for analysis to fix and balance the tower assembly robot at three points. Then, the geometric features of the bolts are extracted based on the acquired image data to identify the bolt posture. Finally, the optimized bolt posture and the bolt tightening scheme with adaptive impedance control are combined to control the tower assembly robot to tighten the bolts to be tightened, accurately tightening the bolts on the transmission tower, thus solving the problem of continuous and significant changes in the relative position between the tower assembly robot and the transmission tower. 2. The present invention innovatively proposes a three-point fixation and balance control technology for a tower assembly robot that integrates multiple sensors. By analyzing image and force sensor data, it generates a fixation and balance strategy for the tower assembly robot and the tower, enabling the robot to be fixed on the working surface and the tower, solving the problem of unequal forces on the traction ropes on both sides of the truss affecting overall stability and safety.

[0025] 3. The present invention innovatively proposes a bolt posture positioning technology. By extracting the geometric features of the bolts, the initial posture of the bolts is estimated based on the extracted geometric features of the bolts. The initial posture is optimized by combining the posture optimization strategy to obtain the optimized bolt posture. This technology can realize multi-type and multi-angle tower bolt positioning based on the bolt geometric features in complex scenarios, thereby improving the accuracy of positioning and identification between the tower assembly robot and the transmission tower.

[0026] 4. The present invention innovatively proposes a bolt tightening control method, combining the robot's coarse-grained working position and the adaptive impedance-controlled bolt tightening scheme to control the tower assembly robot to tighten the bolts to be tightened. This solves the problem of the existing technology that it is difficult to accurately tighten bolts on the transmission tower, and improves the accuracy of positioning and identification between the tower assembly robot and the transmission tower and the efficiency of bolt tightening.

[0027] 5. The present invention innovatively proposes a tower assembly robot, designs a hoisting tower assembly technology based on a hoisting tower assembly robot, and develops a hoisting tower assembly robot. The tower assembly robot is fixed on the transmission tower through a fixing mechanism, and drives the fastening tool to move to the position of the bolt to be fastened through a sliding shaft sliding on the frame, and tightens the bolts to complete the tower assembly task. This solves the problems in the existing technology that climbing tower assembly robots require the design of complex obstacle-crossing structures and are difficult to operate on the tower for a long time, thereby improving the tower assembly efficiency of high-voltage transmission towers.

[0028] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0030] Figure 1 A flow chart of a tower assembly robot control method according to an embodiment of the present invention; Figure 2 A schematic diagram of a tower assembly robot in operation according to an embodiment of the present invention; Figure 3 A schematic structural diagram of a tower assembly robot provided in an embodiment of the present invention; Figure 4 A schematic structural diagram of a bolt fastening mechanism provided in an embodiment of the present invention; Figure 5 An enlarged schematic diagram of a fastening tool portion of a bolt fastening mechanism provided in an embodiment of the present invention; Figure 6 A schematic diagram of an assembly setup of two tower assembly robots provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the overall structure of the tower assembly system provided in an embodiment of the present invention.

[0031] Among them, 1. Tower assembly robot; 111. Frame; 112. Extension; 113. Hoisting mechanism; 114. Lifting ring; 115. Traction rope; 116. Main beam; 12. Fixing mechanism; 131. Sliding shaft; 1311. Longitudinal axis; 1312. Horizontal axis; 1313. Base; 1314. Sliding bracket; 1315. First motor; 1316. Second motor; 132. Fastening tool; 1321. First fastening tool; 1322. Second fastening tool; 2. Transmission tower; 3. Hoisting equipment. DETAILED DESCRIPTION

[0032] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0033] Unless otherwise defined, the technical or scientific terms used herein shall have the ordinary meaning understood by persons of ordinary skill in the field to which the present disclosure belongs. The words “first”, “second”, “third” and similar terms used in the patent application specification and claims of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as “a” or “an” do not indicate a quantity limitation, but rather indicate the presence of at least one. Words such as “include” or “comprise” mean that the elements or objects appearing before “include” or “comprises” include the elements or objects listed after “include” or “comprises” and their equivalents, and do not exclude other elements or objects. Words such as “connect” or “connected” are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Up”, “down”, “left”, “right” and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0034] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0035] As described in the background technology, the tower structure itself is complex. When climbing, the robot needs to cross angle steels in different positions and directions, which requires a high level of obstacle-crossing capability of the climbing mechanism. In addition, the entire climbing process consumes a lot of power, which limits the robot's operating time on the tower, thereby affecting the construction progress. At the same time, during the hoisting process, the tower assembly robot is moved to the set position on the high-voltage transmission tower by hoisting and fixed, and then the bolts are tightened. Because the relative position between the tower assembly robot and the transmission tower will continue to change significantly, it is difficult for the tower assembly robot to accurately tighten the bolts on the transmission tower. An embodiment of the present invention first provides a tower assembly robot control method. Since the relative position between the tower assembly robot and the transmission tower may continue to change significantly, that is, the relative position between the zoom camera and the area to be identified will change significantly, in order to improve the accuracy of recognition and control, a dynamic closed-loop visual servo system is constructed by deeply integrating the multi-scale visual characteristics of the zoom camera with the robot motion control. Its core process is divided into three stages: multi-scale target detection, adaptive adjustment of zoom parameters, and kinematic solution servo control. Each link is tightly coupled to achieve high-precision and high-robustness clamping operations.

[0036] The specific tower assembly robot tower assembly and bolt tightening process is as follows: First, the tower-building robot is hoisted to its operating position using a lifting device. Then, a gripper visual servo guidance algorithm based on zoom-length fusion information is used to obtain image errors and control the motion of the four axes of the tower-building robot's fixed mechanism. A dynamic closed-loop visual servo system is constructed, enabling the fixed mechanism to embrace the angle steel and firmly clamp the tower. A single lifting point and two fixed mechanisms complete the three-point positioning of the tower-building robot. After the tower robot completes its three-point positioning, the traction ropes on both sides of the truss may experience unequal forces, affecting overall stability and safety. By analyzing the force sensor data at the robot's preset positions, a balancing strategy for the robot and the tower is generated to achieve balance between the robot and the tower. After the tower assembly robot firmly clamps the tower body, it adopts a multi-point bolt positioning strategy to locate the bolts. First, data collection is performed, and bolt detection, feature extraction and multi-view feature matching are performed on the collected data. Then, initial pose estimation and BA optimization are performed to output high-precision bolt pose. Bolt tightening is then performed based on a bolt tightening scheme based on adaptive impedance control. High-precision and compliant operation is achieved by combining dynamic parameter adjustment with torque feedback: by dynamically adjusting the impedance parameters (stiffness, damping) of the robot end, the system can adapt to changes in environmental contact force during the bolt screwing process to achieve torque tracking and compliant control. The above method is used in sequence to tighten all the bolts to be tightened. Figure 1 As shown, the specific steps include: Step 1: Analyze the image and force sensor data at the preset positions of the tower assembly robot based on the three-point fixation and balance control method of the tower assembly robot, and generate a three-point fixation and balance strategy for the tower assembly robot and the tower to achieve the fixation and balance of the tower assembly robot on the working surface and the tower; The specific steps include: Step 101: Analyze the image of the tower assembly robot at the preset position based on the three-point fixation method of the tower assembly robot, and generate a three-point fixation strategy for the tower assembly robot and the tower to achieve the fixation of the tower assembly robot on the working surface and the tower, including: identifying a target clamping area for a first image acquired at a first focal length; A focal length optimization objective function is constructed by maximizing the image average gradient and minimizing the image depth estimation error and the focal length variation error, thereby obtaining a second focal length, and acquiring a second image of the target clamping area captured at the second focal length; The weight of the target image at the current second focal length is calculated to perform weighted fusion of the first image and the second image. The target clamping area in the obtained fused image is used to identify the clamping point to control the clamping action of the gripper on the tower assembly robot on the tower body.

[0037] First, the zoom camera is adjusted to the first focal length (short focus) state for short-focus global coarse positioning. At this time, the zoom camera has a first field of view. The target detection model is applied to the collected tower image to detect the target clamping area in the tower image and obtain the target bounding box and the corresponding confidence level.

[0038] Then, when the confidence level meets the requirements, the target bounding box area can be subjected to telephoto local precision detection. At this time, the pan / tilt angle of the zoom camera is adjusted according to the position of the target bounding box, so that the zoom camera is aligned with the center of the target bounding box and switched to the second focal length (telephoto). The second focal length is greater than the first focal length. At this time, the camera has a second field of view angle, which is smaller than the first field of view angle.

[0039] Finally, the area of ​​interest within the zoomed field of view is cropped, segmented, and recognized with high precision to obtain the key points to be clamped by the gripper.

[0040] Specifically, in some embodiments, short focus global coarse positioning: call short focus, such as short focus focal length f s =8mm (FOV 80 degrees), and run the lightweight YOLOv8 model at a set frame rate (30fps in this example). The model input resolution is set to 640×480, and the SPPF module in the YOLOv8 model is used to accelerate feature extraction and output the target bounding box B coarse =[x c ,y c ,w,h] and confidence C, and the target bounding box area is used as the ROI area. When C ≥ 0.85, the telephoto switching condition is triggered.

[0041] Telephoto local precision detection: According to B coarse The center of mass coordinates (x c ,y c ) drives the gimbal to adjust the viewing angle and switch to telephoto, such as the telephoto focal length f l =50mm (12-degree field of view). The ROI area is cropped to 300×300 pixels and then run through the HRNet segmentation network. The network uses a cascaded hourglass architecture, fusing multi-scale features across four resolution branches to output sub-pixel edge keypoints with a positioning accuracy of ±0.1px. The obtained edge keypoints are then used to determine the gripping keypoints of the gripper and control its gripping motion.

[0042] It can be understood that the specific identification process of identifying the target clamping area for images collected in the short-focus state and the specific identification process of identifying the point to be clamped for images collected in the long-focus state can adopt a conventional algorithm model, as long as the target recognition can be achieved. Again, no elaboration is made and no specific limitation is made.

[0043] In this embodiment, in order to ensure image clarity during the zooming process, the zoom parameters are adaptively adjusted to achieve a dynamic trade-off between resolution and field of view.

[0044] When acquiring an image captured at a second focal length, the method includes constructing a focal length optimization objective function by maximizing an average image gradient and minimizing an image depth estimation error and a focal length variation error, thereby obtaining an optimal focal length, and then acquiring a second image captured at the second focal length of the target clamping area; Specifically, the focal length optimization objective function is constructed : ; in, 、 and is the weight coefficient (the default setting is 0.6, 0.3 and 0.1); At the current focal length The depth estimation value of the image under is the mean depth of historical images. It is understandable that the image depth can be estimated using conventional algorithms; The image average gradient is used to define the clarity index; and The maximum focal length and minimum focal length are set; is the set center focal length.

[0045] In this embodiment, the focal length optimization objective function is It contains three main parts: (1) Image clarity part: by maximizing the average gradient of the image, that is, minimizing , optimize image clarity. (2) Estimation consistency part: by minimizing the image depth estimation error , maintaining the consistency of the estimated value. (3) Range constraint part: by minimizing the focal length change error , ensure focal length Within a reasonable range and as close to the center value as possible.

[0046] In this embodiment, the image average gradient for: Define the regional image gradient modulus: ; For the ROI area, the image average gradient As a clarity indicator, The larger it is, the clearer the image; ; in, is the number of pixels selected in the image; Represents the gradient operator, which calculates the image at a certain pixel point The gradient indicates the degree of change in the pixel intensity around the point. The larger the gradient, the more drastic the image change at that point, which also means that there may be edges or details at that point. Indicates the pixels The image gradient modulus at , that is, the rate of change or edge information of the image, measures the intensity change of the image at a certain pixel point and reflects the sharpness of the pixel; and Pixel The horizontal and vertical coordinates of the location.

[0047] In this embodiment, the focal length optimization objective function is solved iteratively by the Levenberg-Marquardt algorithm, and the calculation time of each frame is controlled within 5ms to adjust the focal length in real time to ensure that the image clarity of the target area meets the servo control requirements; at the same time, in the solution process, is greater than the set threshold (such as 25) as a constraint condition, when When it is greater than the set threshold, it means that the clarity meets the servo control requirements. Otherwise, the clarity is insufficient and the focal length optimization objective function is used for re-optimization.

[0048] In this embodiment, in the above zoom process, a dynamic fusion strategy is introduced, and the Sigmoid function is used to achieve a smooth transition of long / short focal length data, thereby constructing a weight function : ; in, Indicates focal length The weight value under is used to adjust the degree of fusion of the long-focus image and the short-focus image; k is the parameter of the Sigmoid function, which controls the speed of the transition. Setting k=0.2 can make the transition smoother and avoid too fast a transition; and For normalized focal length , to make a smooth transition within a certain range.

[0049] Set k=0.2 to balance switching speed and stability, and finally combine the fusion feature map of the long-focus image and the short-focus image for: ; in, To perform bicubic interpolation of short-focus images to long-focus resolution; is the telephoto image feature map (low-resolution image), is the telephoto image weight; is the short-focus image feature map (high-resolution image), through Weighted, and before fusion The function performs upsampling.

[0050] Kinematic solution servo control: When controlling the gripper for gripping, this embodiment employs combined zoom and motion control to address dynamic scenarios. This involves introducing the dynamic focal length term as a variable into the classic image Jacobian matrix via a modified Jacobian matrix to accommodate dynamic zoom lens adjustments in real time. When switching from short focal length (8mm) to long focal length (50mm), the f-value in the matrix increases, and the system automatically adjusts the weight of the rotational component to avoid control instability caused by a reduced field of view.

[0051] Specifically, the improved Jacobian matrix introduces a zoom dynamic term, which enables the gripper control system to adapt to the adjustment of the zoom lens in real time, ensuring that the gripper can still perform stable and effective gripping operations at different focal lengths. The improved Jacobian matrix is ​​expressed as: ; in, Represents the focal length that changes with time t; its role in control: the change of focal length will directly affect the geometric changes of the image, such as the change of field of view. As the focal length increases, the field of view decreases, the image resolution changes, and affects the weights in the Jacobian matrix. The change of focal length affects the rotation and displacement control through the various terms in the matrix.

[0052] Z is the depth of the object, usually indicating how far the object is from the camera. Its role in control: Depth Z affects the relationship between image coordinates and physical space coordinates. Since depth affects changes in perspective, all Z-related terms in the Jacobian matrix involve the adjustment of depth to the control results.

[0053] and The horizontal pixel coordinates and vertical pixel coordinates in the image coordinate system usually represent the coordinates of a point in the image; its role in control: it is used to describe the motion in the image space and affects the accuracy of the grasping operation. The Jacobian matrix is ​​obtained by and The change of is used to calculate the change of image features (such as edges, positions, etc.).

[0054] 、 Calculates the Euclidean distance of a point in the image coordinate system, including the focal length and the sum of the squares of the horizontal and vertical pixel coordinates; Role in control: Reflects geometric changes in the image, especially when the position of the object in the field of view changes, affecting rotation and displacement control.

[0055] Indicates focal length The time rate of change of , i.e., the derivative of focal length. Its role in control: It is used to describe the dynamic effects of focal length changes on the system, especially during zooming, where rapid changes in focal length can affect the controller's response.

[0056] Step 102: Analyze the force sensor data at the preset position of the tower assembly robot based on the tower assembly robot balance control method, and generate a tower assembly robot and tower balance strategy to achieve the balance of the tower assembly robot on the working surface and the tower after being fixed, including: Before the tower-building robot 1 is secured to the transmission tower 2, the lifting ring 114 is located in the middle of the main beam 116, and the traction ropes 115 on both sides of the truss are symmetrical and equally stressed. However, when the first and second fixing structures of the fixing mechanism 12 are grasping angle steel, especially when grasping two angle steels at different positions or inclinations, the main beam 116 is not arranged horizontally after the tower-building robot 1 is secured to the transmission tower 2 via the two fixing mechanisms 12. At this time, the traction ropes 115 on both sides of the truss are subjected to unequal forces. This makes the lifting ring 114 position before the tower-building robot 1 is secured to the transmission tower 2 no longer meet the requirements, which not only affects the overall stability but also damages the traction ropes 115, which are subject to greater forces, affecting stability and safety.

[0057] Based on this, in this embodiment, force sensors are installed on the four traction ropes 115, and force sensors can be set at the connection between the traction ropes 115 and the lifting rings 114; Optionally, a first force sensor and a second force sensor are respectively provided on the two traction ropes 115 at one end of the main beam 116 , and a third force sensor and a fourth force sensor are respectively provided on the two traction ropes 115 at the other end of the main beam 116 .

[0058] After the tower-assembling robot 1 is fixed to the transmission tower 3 by two fixing mechanisms 12, the force average value of the detection values ​​of the first force sensor and the second force sensor is compared with the force average value of the detection values ​​of the third force sensor and the fourth force sensor, and the lifting ring 114 is adjusted to the side with the larger force average value by a lifting machine until the force average values ​​at both ends are equal or the difference between the force average values ​​at both ends is less than a preset difference, thereby avoiding the problem of damage to the traction rope 115 due to a larger force on one end, and at the same time ensuring the overall stability of the tower-assembling robot 1.

[0059] At the same time, the detection value of the first force sensor and the detection value of the second force sensor, as well as the detection value of the third force sensor and the detection value of the fourth force sensor are judged, and the lifting ring 114 is adjusted to the side with the larger detection value through the lifting machine until the two compared detection values ​​are equal or the value difference is less than the preset detection value difference, thereby avoiding the problem of damage to the traction rope 115 due to the greater force on a certain traction rope 115, and at the same time ensuring the overall stability of the tower assembly robot 1.

[0060] Step 2: Extract the geometric features of the corresponding bolts based on the images obtained after the tower assembly robot and the tower are fixed. Combined with the multi-type and multi-angle tower bolt initial positioning method based on the bolt geometric features, the coarse-grained working position of the robot is determined. The tower assembly robot firmly clamps the tower body and adopts a multi-point bolt positioning strategy to locate the bolts. During the bolt positioning process, the tower body image can be captured by a camera to identify the position of the bolt to be tightened. After the positioning of the bolt to be tightened is completed, the position of the bolt to be tightened and the position of the clamping jaw are known, thereby determining the position of the clamping jaw relative to the bolt to be tightened. Then, based on the position of the bolt tightening device relative to the clamping jaw, the position of the bolt tightening device relative to the bolt to be tightened is obtained, thereby controlling the bolt tightening device to move to the position of the bolt to be tightened and controlling the bolt tightening device to perform the tightening operation.

[0061] The specific steps include: Step 201: performing bolt position detection and feature extraction based on the acquired image data; In this embodiment, the camera intrinsic parameters (focal length, distortion parameters) and extrinsic parameters (relative positions of multiple cameras or robot arm motion trajectory) are pre-calibrated. The RGB-D data of the bolt is collected by moving the camera to ensure that adjacent view angles have sufficient overlapping areas.

[0062] In this embodiment, when performing bolt detection and feature extraction, Mask R-CNN is used to detect the instance segmentation mask of the bolt, and the 3D coordinates of the bolt center point are calculated in combination with the depth map. The geometric feature points of the bolt head are extracted, and SuperPoint is used to extract the detail features. Step 202: Initial position estimation; In this embodiment, for each viewpoint, the initial position of the bolt is calculated using the PnP algorithm (solvePnP) by combining 2D key points and the 3D model. Using the 3D point cloud observed from multiple viewpoints, the point cloud is aligned using ICP (Iterative Closest Point) to optimize the initial position. Step 203: Based on the initial position, further optimize and output a high-precision bolt position; In this embodiment, the BA optimization algorithm is combined to optimize the camera pose and the coordinates of the 3D points by minimizing the reprojection error.

[0063] The specific optimization objective function is: , in, is the rotation matrix, is the translation vector of the camera movement, is a robust kernel function to suppress outliers, is the camera projection model, is the 3D coordinate of the jth key point of the bolt, is the 2D observation of the jth key point in view i; Finally, the coordinates of the 3D points of the optimized bolt are obtained; It is understandable that after the bolt to be tightened is positioned, the position of the bolt to be tightened and the position of the clamp are known, thereby determining the position of the clamp relative to the bolt to be tightened. Then, based on the position of the bolt tightening device relative to the clamp, the position of the bolt tightening device relative to the bolt to be tightened is determined. This is used to control the bolt tightening device to move to the position of the bolt to be tightened and control the bolt tightening device to perform the tightening operation. This invention proposes a multi-type and multi-angle tower bolt positioning technology based on the geometric characteristics of the bolts in complex scenarios, which realizes the initial positioning of the tower tightening operation position by the robot.

[0064] Step 3: Based on the coarse-grained working position of the robot, combined with the bolt and tool matching method of adaptive impedance control, generate the robot's fine-grained working position corresponding to the operation of multiple bolts and tool matching to control the tower assembly robot to tighten the bolts to be tightened.

[0065] After obtaining the coarse-grained working position of the robot, the control of the robot arm's posture, angle, displacement and other motion processes is designed to move the robot arm into position. When the robot arm is linked to tighten bolts, a bolt tightening solution based on adaptive impedance control is used for bolt tightening. The specific steps include the following: Step 301: constructing an expected impedance equation; In this embodiment, the expected impedance equation is:

[0066] in, 、 and denote the target inertia, damping and stiffness matrices respectively; is the desired position of the robotic arm, is the actual position of the robotic arm, is the end contact force, by adjusting and , so that the actual contact force Keep track of the desired force, which can be determined based on the bolt size.

[0067] Step 302: updating the stiffness of the environment in real time to capture the stiffness changes during the bolt screwing phase, such as thread engagement and locking; In this embodiment, the estimation of the stiffness of the environment can be achieved by using a recursive least squares method, which is expressed as:

[0068] in, represents the environmental stiffness, Indicates the contact point position between the robot arm and the bolt, represents the updated environmental stiffness; Step 303: updating the expected impedance equation based on the damping adjustment rule and the stiffness adjustment rule to obtain an updated expected impedance equation; Specifically, the stiffness adjustment rule and the damping adjustment rule and include: , in, represents the initial stiffness, Represents the adaptive gain coefficient, when When the force is less than the expected value, increase To enhance the response, Greater than expected When Avoid overloading; , in, represents the initial damping, Represents the damping attenuation coefficient, which dynamically adjusts the damping according to the torque change rate to suppress oscillation; Step 304: Correct the position based on the updated expected impedance equation and the environmental stiffness to obtain the expected position, thereby controlling the tower assembly robot to tighten the bolts to be tightened. The correction to the desired position is expressed as: ; Combining dynamic parameter adjustment with torque feedback enables high-precision, compliant operation: By dynamically adjusting the impedance parameters (stiffness and damping) of the robot's end effector, the system adapts to changes in the ambient contact force during bolt insertion, achieving torque tracking and compliant control. This method is then applied sequentially to all bolts to be tightened.

[0069] Based on the above scheme, the control of the tower assembly robot from clamping, bolt positioning to bolt tightening is realized, and the tower assembly robot can tighten the fastening bolts in sequence during the tower assembly operation, thereby meeting the tower assembly requirements.

[0070] In some embodiments, a tower assembly robot control system is further provided, comprising: A fixation and balance control module is used to analyze the image and force sensor data at the preset position of the tower assembly robot based on the three-point fixation and balance control method of the tower assembly robot, and generate a three-point fixation and balance strategy for the tower assembly robot and the tower to achieve fixation and balance of the tower assembly robot on the working surface and the tower; The bolt pose optimization module is used to extract the corresponding bolt geometric features based on the images obtained after the tower assembly robot and the tower are fixed. It combines the multi-type and multi-angle tower bolt initial positioning method with the bolt geometric features to determine the coarse-grained working position of the robot; The tightening control module is used to generate multiple bolt operations and the corresponding fine-grained robot operation positions based on the robot's coarse-grained operation positions and the bolt and tool matching method with adaptive impedance control to control the tower assembly robot to tighten the bolts to be tightened.

[0071] It should be noted that the specific implementation method of a tower assembly robot control system in an embodiment of the present invention is similar to the specific implementation method of a tower assembly robot control method in an embodiment of the present invention. Please refer to the description of the method part for details. In order to reduce redundancy, it will not be repeated here.

[0072] The embodiment of the present invention further provides a tower assembly robot, such as Figure 2-Figure 6 As shown, it includes: a robot body frame, a fixing mechanism 12, a bolt fastening mechanism, a zoom camera and a control module; The robot body frame includes a frame 111 and a lifting mechanism 113 connected to the upper part of the frame 111; The fixing mechanism 12 is provided at the bottom of the frame 111 and is used to fix the robot body frame to the tower body of the transmission tower 2; The bolt fastening mechanism includes a sliding shaft 131 and a fastening tool 132. One end of the sliding shaft 131 is slidably connected to the frame 111, and the other end is rotatably connected to the fastening tool 132. The sliding shaft is used to change the relative position of the fastening tool and the transmission tower, and the fastening tool is used to tighten the bolts on the transmission tower to perform tower assembly operations. The zoom camera is arranged at the fixing mechanism 12, and the control module is arranged on the robot body frame. The zoom camera is used to obtain images collected at a first focal length and a second focal length; The control module is configured to execute the tower assembly robot control method described in the above embodiment. For details, please refer to the description of the method part. In order to reduce redundancy, it is not repeated here.

[0073] This embodiment addresses the problems existing in existing climbing tower assembly robots and adopts the concept of hoisting operations to design a hoisting tower assembly robot. Compared to climbing tower assembly robots, there is no need to design additional complex climbing structures, and it can more flexibly perform assembly operations on various types of transmission towers, saving time and energy consumption during the robot climbing process, thereby improving work efficiency. During tower assembly operations, the robot is hoisted to a set position on the transmission tower, then fixed to the tower body by a fixing mechanism, and then the fastening tool is driven by a sliding shaft to move to the position of the bolt to be fastened for bolt tightening. At the same time, the bolt posture is extracted based on the fused image, and the optimized bolt posture and the bolt tightening scheme with adaptive impedance control are combined to control the tower assembly robot to tighten the bolt to be fastened. This solves the problem of the difficulty in accurately fastening bolts on the transmission tower in the existing technology, improves the accuracy of positioning and identification between the tower assembly robot and the transmission tower, and improves the efficiency of bolt tightening.

[0074] Specifically, the frame 111 includes a horizontal main beam 116 and extensions 112 at each end of the main beam. The extensions 112 extend upward to a set height above the main beam. A lifting mechanism 113 has a lifting ring 114 at its top. Four traction ropes 115 connect the ends of the lifting mechanism to the tops of the extensions at each end of the main beam to meet lifting requirements.

[0075] The sliding shaft 131 has three degrees of freedom of movement in the X, Y, and Z directions. At the same time, the fastening tool is rotatably connected to the end of the sliding shaft 131 away from the frame, providing rotational freedom to meet the bolt tightening requirements of the fastening tool 132 at any position. In this embodiment, the sliding shaft includes a transverse axis 1312 and a longitudinal axis 1311. Both the transverse axis 1312 and the longitudinal axis 1311 can adopt existing linear guides or other structures that can meet the requirements of linear motion. The longitudinal axis 1311 is vertically arranged and perpendicular to the frame 111. One end of the longitudinal axis 1311 is slidably connected to the frame 111, and the other end is vertically upward; the transverse axis 1312 is slidably connected to the end of the longitudinal axis 1311 away from the frame 111, and the transverse axis and the longitudinal axis are perpendicular to each other. The longitudinal axis 1311 has a first slider that can slide along the length of the longitudinal axis. The transverse axis 1312 is fixedly connected to the first slider, so that it can move vertically along the longitudinal axis under the drive of the first slider.

[0076] A base 1313 is provided at one end of the longitudinal shaft 1311, which is slidably connected to the frame 111. Base 1313 is constructed from two perpendicular plates fixedly connected, with reinforcing ribs interlocking the two plates. One plate is fixedly connected to the longitudinal shaft 1311, while the other plate is slidably connected to the frame. A sliding rail is provided on the upper portion of the frame 111, on which a third slider is slidably mounted. Another plate is fixedly connected to the third slider, thereby achieving the sliding connection between the longitudinal shaft and the frame. The method for achieving linear motion by cooperating between the sliding rail and the slider can be achieved using existing technology and will not be further described here.

[0077] In some embodiments, the fastening tool 132 is connected to the cross shaft 1312 through a sliding bracket 1314. The first end of the sliding bracket 1314 is slidably connected to the end of the cross shaft far from the longitudinal axis, and the second end of the sliding bracket 1314 is rotatably connected to the fastening tool. To achieve more flexible movements and meet the bolt fastening requirements at different positions, the fastening tool 132 has not only a rotational freedom degree but also a sliding freedom degree with respect to the cross shaft. There is a second slider on the cross shaft 1312 that can slide along the length direction of the cross shaft. The sliding bracket is of a U-shaped structure, with one end fixedly connected to the second slider and the other end connected to the fastening tool 132 through a rotational connection structure, so as to move along the length direction of the cross shaft under the drive of the second slider and rotate relative to the cross shaft.

[0078] To meet the fastening requirements of different tower body types and angles at different directions and positions, in this embodiment, between the fastening tool and the sliding bracket, there are at least two rotational freedom degrees of horizontal rotation and vertical rotation. Specifically, the rotation control of the fastening tool is achieved through a rotation motor. The body of the first motor 1315 is fixedly connected to the sliding bracket 1314, and the output end is connected to the second motor 1316. The output end of the second motor 1316 is further connected to the fastening tool through a bracket. The output ends of the first motor 1315 and the second motor 1316 are perpendicular to each other, and the rotation axis of the output end of the first motor 1316 is in the vertical direction, and the rotation axis of the output end of the second motor 1316 is in the horizontal direction, thus meeting the requirement of having two rotational freedom degrees in the horizontal and vertical directions.

[0079] In some embodiments, to further improve the operation efficiency, the number of the fastening tools 132 can be set to be multiple, and each fastening tool can rotate independently relative to the sliding bracket 1314, so as to be able to fasten bolts at different positions separately or simultaneously. In this embodiment, the number of the fastening tools 132 is 2. The first fastening tool 1321 and the second fastening tool 1322 are both connected to the output end of the second motor 1316 through a bracket, and the connection positions of the first fastening tool 1321 and the second fastening tool 1322 with the bracket can rotate to achieve independent adjustment and control. Based on this concept, the number of the fastening tools can be further increased to further meet the simultaneous fastening requirements of bolts at different positions. For example, it can achieve simultaneous fastening of both ends of an angle steel, thus avoiding the dislocation of the angle steel during the fastening process.

[0080] The fixing mechanism 12 includes a first fixing mechanism and a second fixing mechanism. The first fixing mechanism and the second fixing mechanism have the same structure and are respectively arranged at the two ends of the bottom of the frame and can slide relative to the frame. The two fixing mechanisms cooperate with the frame in the same way. Taking the first fixing mechanism as an example, in some embodiments, a slide rail is provided at one end of the bottom of the frame 111, and a fourth slider is slidably provided on the slide rail. The first fixing mechanism includes a clamping claw and a first linear guide rail and a second linear guide rail that are perpendicular to each other. The first linear guide rail is fixedly connected to the fourth slider. The first linear guide rail has a fifth slider that can slide along the length direction of the first linear guide rail. The second linear guide rail is fixedly connected to the fifth slider, and the second linear guide rail has a sixth slider that can slide along the length direction of the second linear guide rail. The clamping claw is connected to the sixth slider, so that it can move along the X, Y, and Z directions under the sliding drive of the fourth slider, the fifth slider, and the sixth slider, thereby clamping the tower body at an appropriate position to provide a support point for the robot.

[0081] In some embodiments, a tower assembly system is further provided, such as Figure 1 As shown, the tower erection robot 1 provided above is provided with a hoisting device 3. The hoisting device 3 has a lifting arm that can cooperate with the lifting mechanism of the tower erection robot 1. The hoisting device can be a crane or other engineering machinery with a lifting function, or an aerial work platform equipped with a hook, as long as it has a hook and can set the height of the tower erection robot's hoist device.

[0082] To further improve tower assembly efficiency, and considering the different sizes (mainly widths) of different tower types (such as wine glass-shaped towers, cathead-shaped towers, and dry-shaped towers), a tower assembly robot with a fixed width may be difficult to fix at a wider position on the transmission tower. The tower assembly robot can also be configured to be spliced ​​together in a form where the racks of multiple tower assembly robots are fixedly connected in the length direction, thereby increasing the working area and providing support in multiple positions. Take two tower assembly robots assembled together as an example. Figure 6 As shown, the extension parts of adjacent tower-assembling robots can be fixedly connected together through a connecting structure, and the two tower-assembling robots can be assembled into a whole, and then hoisted as a whole to perform tower assembly operations.

[0083] In some embodiments, based on the tower assembly robot or tower assembly system provided above, a tower assembly method is further provided, comprising: Step 1: Hoist the tower assembly robot 1 to the set position of the transmission tower 2; Step 2: Use the tower assembly robot's fixing mechanism to fix the tower assembly robot on the transmission tower; three-point positioning is formed by the lifting point and two fixing mechanisms; Step 3: The sliding shaft moves, driving the fastening tool to the position of the bolt to be fastened on the transmission tower; Step 4: The tightening tool moves to tighten the bolts to be tightened.

[0084] When tower assembly is required, the tower assembly robot is hoisted to a desired height and brought close to the transmission tower using a hook suspended from an external lifting device (such as a crane). The robot then uses a zoom camera mounted on its gripper 1 to identify the tower structure within its field of view, determine the appropriate gripper position, and then control the gripper to extend and secure the gripper at the designated location. The robot has at least two grippers, which, in conjunction with the hook of the lifting device, provide three-point positioning for the tower assembly robot, accurately and securely securing it to the transmission tower. The robot's bolt tightening device 2 then activates, identifying the bolts to be tightened on the tower and tightening them, completing the tower assembly.

[0085] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A tower assembly robot control method, characterized in that: The steps include: Based on the tower assembly robot's three-point fixation and balance control method, the image and force sensor data at the preset position of the tower assembly robot are analyzed to generate a three-point fixation and balance strategy for the tower assembly robot and the tower, so as to achieve the fixation and balance of the tower assembly robot on the working surface and the tower. Based on the images obtained after the tower assembly robot and the tower are fixed, the geometric features of the corresponding bolts are extracted. Combined with the multi-type and multi-angle tower bolt initial positioning method based on the bolt geometric features, the coarse-grained working position of the robot is determined. Based on the coarse-grained working position of the robot and combined with the bolt and tool matching method of adaptive impedance control, the operations of multiple bolts and the corresponding fine-grained working positions of the robot are generated to control the tower assembly robot to tighten the bolts to be tightened.

2. A tower assembly robot control method according to claim 1, characterized in that: Based on the three-point fixation method of the tower assembly robot, the image at the preset position of the tower assembly robot is analyzed to generate a three-point fixation strategy for the tower assembly robot and the tower to achieve the fixation of the tower assembly robot on the working surface and the tower, including: identifying a target clamping area for a first image acquired at a first focal length; A focal length optimization objective function is constructed by maximizing the image average gradient and minimizing the image depth estimation error and the focal length variation error. After obtaining the optimal focal length, a second image of the target clamping area is acquired at the optimal focal length. The weight of the target image at the current optimal focal length is calculated, and the first image and the second image are weightedly fused. The target clamping area in the obtained fused image is used to identify the clamping point to control the clamping action of the gripper on the tower assembly robot on the tower body.

3. A tower assembly robot control method according to claim 1, characterized in that: Based on the tower assembly robot balance control method, the force sensor data at the preset position of the tower assembly robot is analyzed to generate a tower assembly robot and tower balance strategy to achieve the balance of the tower assembly robot after the working surface and the tower are fixed, including: The average force value detected by the first force sensor and the second force sensor is compared with the average force value detected by the third force sensor and the fourth force sensor, and the lifting ring is adjusted toward the side with the larger average force value by the lifting machine until the average forces at both ends are equal or the difference between the average forces at both ends is less than a preset difference; At the same time, the detection value of the first force sensor and the detection value of the second force sensor, as well as the detection value of the third force sensor and the detection value of the fourth force sensor are judged, and the lifting ring is adjusted to the side with the larger detection value through the lifting machine until the two compared detection values ​​are equal or the difference between the values ​​is less than the preset detection value difference.

4. A tower assembly robot control method according to claim 1, characterized in that: The method extracts geometric features of corresponding bolts based on images obtained after the tower assembly robot and the tower are fixed, combines the geometric features of bolts with the multi-type and multi-angle tower bolt initial positioning method to determine the coarse-grained working position of the robot, including: Acquire images captured at the first focal length and the second focal length after the tower assembly robot and the tower are fixed; The images captured at the first focal length and the second focal length are fused to obtain a fused image. The geometric features of the bolt are extracted based on the fused image, and the initial pose of the bolt is estimated based on the extracted geometric features of the bolt. The initial pose is optimized by combining the pose optimization strategy to obtain the optimized bolt pose. The optimized bolt pose is used to control the clamping action of the gripper on the tower assembly robot on the tower body.

5. A tower assembly robot control method according to claim 4, characterized in that: When acquiring an image captured at the second focal length, a focal length optimization objective function is constructed by maximizing the image average gradient and minimizing the image depth estimation error and the focal length change error. After obtaining the optimal focal length, a second image of the target clamping area captured at the second focal length is acquired.

6. A tower assembly robot control method according to claim 5, characterized in that: The focal length optimization objective function is: ； in, 、 and is the weight coefficient; At the current focal length The depth estimation value of the image under is the mean depth of historical images; is the average gradient of the image; and is the maximum focal length and the minimum focal length; is the center focal length.

7. A tower assembly robot control method according to claim 1, characterized in that: The geometric features of the corresponding bolts are extracted based on the images obtained after the tower assembly robot and the tower are fixed. The multi-type and multi-angle tower bolt initial positioning method based on the bolt geometric features is used to determine the coarse-grained working position of the robot, including: extracting the geometric features of the bolts based on the obtained image data; The initial position of the bolt is estimated based on its geometric features. The coordinate position of the 3D point of the bolt is optimized by minimizing the reprojection error based on the BA optimization algorithm. The position of the clamp relative to the bolt to be tightened is determined based on the position of the bolt to be tightened and the position of the clamp. Then, the position of the bolt tightening device relative to the bolt to be tightened is obtained based on the position of the bolt tightening device relative to the clamp.

8. A tower assembly robot control method according to claim 1, characterized in that: The bolt and tool matching method combined with adaptive impedance control generates multiple bolt operations and tool matching corresponding robot fine-grained operation positions, including: Construct the expected impedance equation; The stiffness of the environment is updated in real time to capture the stiffness changes during the bolt screwing phase. The expected impedance equation is updated based on the damping adjustment rule and the stiffness adjustment rule to obtain an updated expected impedance equation. The position is corrected based on the updated expected impedance equation and environmental stiffness to obtain the expected position, thereby controlling the tower assembly robot to tighten the bolts to be tightened.

9. A tower assembly robot control method according to claim 8, characterized in that: Stiffness adjustment rules and damping adjustment rules and include: ， in, represents the initial stiffness, Represents the adaptive gain coefficient, when When the force is less than the expected value, increase To enhance the response, Greater than expected When Avoid overloading; ， in, represents the initial damping, It represents the damping attenuation coefficient, which dynamically adjusts the damping according to the torque change rate to suppress oscillation.

10. A tower assembly robot control system, characterized in that: include: A fixation and balance control module is used to analyze the image and force sensor data at the preset position of the tower assembly robot based on the three-point fixation and balance control method of the tower assembly robot, and generate a three-point fixation and balance strategy for the tower assembly robot and the tower to achieve fixation and balance of the tower assembly robot on the working surface and the tower; The bolt pose optimization module is used to extract the corresponding bolt geometric features based on the images obtained after the tower assembly robot and the tower are fixed. It combines the multi-type and multi-angle tower bolt initial positioning method with the bolt geometric features to determine the coarse-grained working position of the robot; The tightening control module is used to generate multiple bolt operations and the corresponding fine-grained robot operation positions based on the robot's coarse-grained operation positions and the bolt and tool matching method with adaptive impedance control to control the tower assembly robot to tighten the bolts to be tightened.

11. A tower assembly robot, characterized in that: include: Robot body frame, fixing mechanism and bolt fastening mechanism, zoom camera and control module; The robot body frame includes a frame and a lifting mechanism connected to the upper part of the frame; The fixing mechanism is arranged at the bottom of the frame and is used to fix the robot body frame to the tower body of the transmission tower; The bolt fastening mechanism includes a sliding shaft and a fastening tool, one end of the sliding shaft is slidably connected to the frame, and the other end is rotatably connected to the fastening tool, which is used to change the relative position of the fastening tool and the transmission tower, and the transmission tower is assembled by the fastening tool; The zoom camera is arranged at the fixing mechanism, and the control module is arranged on the robot body frame. The zoom camera is used to obtain images collected at a first focal length and a second focal length; The control module is configured to execute the tower assembly robot control method according to any one of claims 1 to 9.

12. The tower assembly robot according to claim 11, characterized in that: The sliding shaft includes a longitudinal shaft and a transverse shaft, one end of the longitudinal shaft is slidably connected to the frame, the transverse shaft is slidably connected to an end of the longitudinal shaft away from the frame, and the transverse shaft is perpendicular to the longitudinal shaft.

13. The tower assembly robot according to claim 12, characterized in that: The fastening tool is connected to the transverse axis via a sliding bracket, a first end of the sliding bracket is slidably connected to an end of the transverse axis away from the longitudinal axis, and a second end of the sliding bracket is rotatably connected to the fastening tool.

14. The tower assembly robot according to claim 13, characterized in that: There are at least two rotational degrees of freedom between the fastening tool and the sliding bracket: horizontal rotation and vertical rotation.

15. The tower assembly robot according to claim 14, characterized in that: There are multiple fastening tools, and each fastening tool can rotate independently relative to the sliding bracket.

16. The tower assembly robot according to claim 11, characterized in that: The fixing mechanism includes a first fixing mechanism and a second fixing mechanism. The first fixing mechanism and the second fixing mechanism are respectively arranged at two ends of the bottom of the frame and can slide relative to the frame.

17. The tower assembly robot according to claim 14, characterized in that: The frame includes a main beam and extensions arranged at both ends of the main beam, the extensions extending upward to a set height higher than the main beam; the top of the lifting mechanism is provided with a lifting ring, and the two ends of the lifting mechanism are respectively connected to the tops of the extensions at both ends of the main beam.

18. A tower assembly system, characterized in that: The tower erection robot comprises a hoisting device and the tower erection robot according to any one of claims 11 to 17, wherein the hoisting device has a hoisting arm capable of cooperating with the hoisting mechanism of the tower erection robot.

19. A tower assembly system according to claim 18, characterized in that: There are multiple tower assembly robots, and the frames of the multiple tower assembly robots are fixedly connected in the length direction.

20. A tower assembly method, characterized in that: Based on the tower assembly robot according to any one of claims 11 to 17, or the tower assembly system according to any one of claims 18 to 19, comprising: Hoist the tower assembly robot to the set position of the transmission tower; Fix the tower assembly robot on the transmission tower through the fixing mechanism of the tower assembly robot; The sliding shaft moves, driving the fastening tool to the position of the bolt to be fastened on the transmission tower; The tightening tool moves to tighten the bolt to be tightened.

Images