Overhead cable multi-agent cooperative anti-icing coating control method and robot

Abstract

The invention belongs to the technical field of power robots. According to the multi-agent cooperative anti-icing coating control method for the overhead cable and the robot, when a lower coating assembly cannot be separated from the overhead cable, or the lower coating assembly cannot move to a set off-line position although the lower coating assembly is separated from the overhead cable, it is judged that the lower coating assembly is in an emergency off-line state; in the emergency off-line state, the emergency lifting appliance is pulled through the unmanned aerial vehicle, and the synchronous belt is driven to act so that the lower coating assembly can move in the direction away from the upper coating assembly; and when the lower coating assembly moves to a set off-line position, emergency off-line is completed. The problem that the anti-icing coating device lacks an emergency off-line structure support is solved, the pertinence and timeliness of emergency response are improved, and the risk that the device is detained in the cable due to the fact that the off-line operation is delayed because the emergency state cannot be judged is avoided.

Description

Technical Field

[0001] This invention relates to the field of power robot technology, and in particular to a multi-agent collaborative anti-icing coating control method and robot for overhead cables. Background Technology

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Overhead power lines, as the core medium for long-distance power transmission, are widely used in power systems. In low-temperature winter conditions, overhead lines are prone to icing, a problem that has become a significant threat to the safe and stable operation of the power grid. This can lead to conductor galloping and, in severe cases, even major ice-related accidents such as tower collapse and line breakage. To address the problem of line icing, the industry has developed two main technical approaches: active de-icing and passive protection. Passive coating, due to its economic and practical advantages, has gradually become one of the mainstream protection methods. Traditional manual climbing and painting methods suffer from poor safety and low efficiency, failing to meet the needs of large-scale line protection. Therefore, the development of anti-icing coating robots has become an urgent need for industry development. Currently, various anti-icing coating devices for transmission lines have emerged in related technical fields. These devices mostly adopt an open-and-close structure design, using nozzles to spray coatings, and are equipped with functional modules such as drying and residual material recovery. Some devices also integrate de-icing mechanisms, achieving integrated de-icing and coating operations, providing diverse technical solutions for line anti-icing.

[0004] Although existing anti-icing coating devices have basic coating functions, existing overhead cable anti-icing coating technology still suffers from low automation, insufficient accuracy in identifying critical conditions, and poor adaptability of process parameters, making it difficult to guarantee the quality and efficiency of the operation. Specifically, the problems are as follows: Existing overhead cable anti-icing coating operations rely heavily on manual labor, lacking automated control throughout the entire process from equipment mounting and coating process management to offline recycling; the reliability of key equipment and cable clamping / loosening status judgments is insufficient, and there is a lack of effective feedback and strategy adjustment mechanisms during the coating process, resulting in low operational efficiency and high safety risks; in overhead cable anti-icing coating operations, the judgment of equipment and cable clamping status lacks accurate and effective technical means, is greatly affected by the working environment, and has low accuracy in recognizing the contact and adhesion status between the clamping wheel and the cable, easily leading to insufficient stability of the equipment after mounting or improper operation when dismounting, affecting the continuity and safety of the operation; in existing coating operations, key parameters such as walking speed and material feeding speed are mostly fixed settings, unable to be dynamically adapted and adjusted according to the actual coating effect, easily resulting in uneven coating thickness; at the same time, the impact of ambient temperature on the acquisition quality is not considered during image acquisition, which interferes with the accuracy of subsequent effect judgment and parameter adjustment, affecting the coating quality. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a multi-agent collaborative anti-icing coating control method and robot for overhead cables. This method solves the problems of traditional overhead cable anti-icing coating relying on manual labor, low automation, and disconnect between status monitoring and strategy adjustment during the coating process. It improves the automation level and efficiency of anti-icing coating operations, enhances the precision control and reliability of the coating process, avoids the safety risks of manual high-altitude operations, and strengthens the adaptability of the equipment to overhead cable operation scenarios.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a multi-agent collaborative anti-icing coating control method for overhead cables.

[0007] A multi-agent collaborative anti-icing coating control method for overhead cables utilizes an anti-icing coating robot including a clamping module, a cleaning module, and a coating module, comprising the following processes: The robot's online clamping status autonomous judgment steps are as follows: After the anti-icing coating robot is hoisted onto the cable, the control clamping module executes the clamping action, and the clamping judgment agent is called. Based on the image detection confidence after the clamping action is completed, the clamping judgment result is generated. When the clamping judgment result is that the robot is clamped, the anti-icing coating robot is online. The robot coating effect dynamic feedback recognition steps are as follows: start the cleaning module to clean the cable, control the anti-icing coating robot to walk along the cable, control the coating module to coat the cleaned cable, call the coating effect recognition intelligent agent, generate the coating effect recognition result based on the image after coating, and call the closed-loop decision intelligent agent to dynamically adjust the coating strategy of the coating module based on the coating effect recognition result. The fully autonomous lifting and unloading control steps of the robot are as follows: After the anti-icing coating robot completes the coating action, the clamping module is controlled to perform the loosening action. The clamping judgment agent is called to generate a loosening judgment result based on the image after the loosening action is completed. When the loosening judgment result is that it has been loosened, the anti-icing coating robot is lifted and unloaded.

[0008] In one implementation of the first aspect of the present invention, in the autonomous judgment step of the robot's online clamping state, a clamping discrimination agent is invoked to generate a clamping judgment result based on the image detection confidence level after the clamping action is completed, including: The image after the clamping action is completed is preprocessed. Based on the preprocessed image, the enhanced feature map is extracted according to the CSPDarknet-53 backbone network of the YOLOv5 network model. The enhanced feature map is used as the input to the detection head of the YOLOv5 network model to detect clamping wheels and cable targets. The output is the target bounding box, confidence score and category. Valid clamping wheel targets with a confidence score not lower than the threshold are selected. and effective cable target ; calculate and Contact overlap Calculate the mean confidence level of the detection. According to the degree of contact overlap With the mean confidence level of the detection The integration yielded a comprehensive score. ; When the overall score If the value is greater than or equal to the set threshold, it is determined to be compressed; otherwise, it is determined to be loosened.

[0009] In one implementation of the first aspect of the present invention, calculation and Contact overlap ,include: ; in, This is the area calculation function.

[0010] In one implementation of the first aspect of the present invention, the mean confidence level is detected. for: , and These represent the effective clamping wheel targets. and effective cable target Confidence level; To detect the mean confidence level With contact overlap The weighted sum is used as the comprehensive score. .

[0011] In one implementation of the first aspect of the present invention, the step of dynamically feeding back and recognizing the coating effect of the robot includes invoking a closed-loop decision-making agent to dynamically adjust the coating strategy of the coating module based on the coating effect recognition result, comprising: The actual coating thickness is determined based on the coating effect recognition results. The walking speed of the anti-icing coating robot is dynamically adjusted based on the target coating thickness and the actual coating thickness, including: ; in, This represents the corrected walking speed; Represents the initial walking speed; Represents the target coating thickness; This represents the actual coating thickness.

[0012] As a further limitation of the first aspect of the invention, the real-time coating supply speed is determined based on the actual diameter of the cable, the walking speed of the anti-icing coating robot, and the target coating thickness, including: ,in, This represents the coating compatibility correction factor. Represents pi (π) Represents the actual outer diameter of the cable. This represents the walking speed of the anti-icing coated robot. Represents the target coating thickness. This represents the real-time feed rate of the coating.

[0013] In one implementation of the first aspect of the present invention, dynamic compensation of exposure time is performed during the image acquisition process after the pressing action is completed, the image acquisition process after the coating is completed, and the image acquisition process after the releasing action is completed, including: ; in, This represents the exposure time after compensation; Represents the baseline exposure time; Represents the temperature compensation coefficient; Represents the reference temperature; This represents the actual ambient temperature.

[0014] In one implementation of the first aspect of the present invention, the coating module includes an upper coating component, a lower coating component, and a frame component. The lower coating component is connected to a synchronous belt on the frame component and can move up and down under the drive of the synchronous belt to cooperate with the upper coating component to cover the overhead cable. An emergency lifting device is connected to the synchronous belt. If the coating component cannot detach from the overhead cable, or if the lower coating component detaches from the overhead cable but cannot move to the designated disconnection position, it is determined to be in an emergency disconnection state. In an emergency disconnection situation, an emergency lifting device is pulled by a drone, which drives the synchronous belt to move the lower coating component away from the upper coating component; Emergency de-corrosion is completed when the coating component moves to the set de-corrosion position.

[0015] As a further limitation of the first aspect of the present invention, in the emergency disconnection state, the controller of the anti-icing coating device automatically sends an emergency disconnection request to the UAV, so that the UAV pulls the emergency lifting device via the suspended boom according to the received request; or, in the emergency disconnection state, the controller of the anti-icing coating device sends an emergency disconnection request to the ground control terminal, so that the UAV pulls the emergency lifting device according to the received request; or, in the emergency disconnection state, the UAV is directly controlled to pull the emergency lifting device via the ground control terminal.

[0016] Secondly, the present invention provides an anti-icing coating robot.

[0017] An anti-icing coating robot, comprising: The robot body and the cleaning module, clamping module, control terminal, walking wheel set, clamping image acquisition unit, coating image acquisition unit and coating module arranged on the robot body; The cleaning module is located at the front end of the robot body, the coating module is located at the rear end of the robot body, the clamping module is located in the middle of the robot body, and the walking wheel set includes a first wheel set and a second wheel set. The first wheel set is located between the cleaning module and the clamping module, and the second wheel set is located between the coating module and the clamping module. The first wheel set and the second wheel set are driven by the walking motor on the robot body. The control terminal is communicatively connected to the cleaning module, coating module, pressing module, control terminal, walking wheel set, pressing image acquisition unit, and coating image acquisition unit.

[0018] In one implementation of the second aspect of the present invention, the clamping module includes: a clamping wheel, a clamping motor, a first position switch and a second position switch. The clamping wheel can move up and down under the drive of the clamping motor to clamp or loosen the cable. The first position switch and the second position switch are arranged vertically. The first position switch is used to identify the clamping state and the second position switch is used to identify the loosening state.

[0019] In one implementation of the second aspect of the present invention, the coating module includes: an upper coating component, a lower coating component, a floating wheel component, and a frame component. The lower coating component is connected to a synchronous belt on the frame component and can move up and down under the drive of the synchronous belt to cooperate with the upper coating component to cover the overhead cable. An emergency lifting device is connected to the timing belt. The emergency lifting device is used to drive the timing belt in case of emergency disconnection, so that the lower coating assembly moves away from the upper coating assembly. The floating wheel assembly includes a support frame and floating wheels arranged on the support frame for traveling on overhead cables. The upper coating assembly is connected to the support frame, which is used to connect to the robot body of the anti-icing coating robot.

[0020] As a further limitation of the second aspect of the present invention, the upper coating assembly includes: a first foam cover plate, a first foam, a first quick-connect connector, and an upper coating bracket. The first foam cover plate is fixed on the upper coating bracket, forming an upper inner cavity with the opening facing downwards. The first foam is fixed inside the upper inner cavity and is used to apply the coating to the overhead cable. One end of the first quick connector is connected to the coating supply system and is fixed on the upper coating bracket and communicates with the upper inner cavity.

[0021] As a further limitation of the second aspect of the present invention, the floating wheel assembly includes: a locking pin, a support frame, a floating wheel, and a driver, wherein the support frame includes a left link, a right link, and a driver fixing frame connected between the left link and the right link, and the driver is fixed on the driver fixing frame; The floating wheel shaft fixes the floating wheel to the bottom of the drive frame, and the support frame is connected to the robot body of the coating robot by locking pins and can slide up and down along the robot body.

[0022] As a further limitation of the second aspect of the present invention, the undercoating assembly includes: a sensor detection sheet, a belt pressure block, a slider fixing plate, a second quick-change connector, an undercoating bracket, a second foam, and a second foam cover plate. The second foam cover plate is fixed on the lower coating bracket to form an upward-opening lower inner cavity. The second foam is fixed inside the lower inner cavity and is used to apply the coating to the overhead cable. One end of the second quick connector is connected to the paint recovery system, and the other end of the second quick connector is fixed to the lower coating bracket and communicates with the lower inner cavity. One side of the slider fixing plate is fixed to the belt pressure block with bolts, and the other side of the slider fixing plate is fixedly connected to the lower coating bracket. The sensor detection plate is fixed to one side of the slider fixing plate and is used to detect the position of the lower coating component. A slider is connected to the slider fixing plate.

[0023] As a further limitation of the second aspect of the present invention, the frame assembly includes: a motor mounting plate, a linear guide rail, a synchronous belt, an idler wheel, an idler wheel shaft, an adjusting bolt mounting plate, an adjusting bolt, a synchronous belt pulley, a photoelectric switch, a drive motor, a guide rail pad, an idler wheel mounting bracket, and a mounting base plate. The drive motor is fixed to the motor mounting plate by bolts, and the motor mounting plate is fixed to the base plate. The synchronous pulley is fixed on the output shaft of the drive motor. The synchronous pulley is connected to the idler pulley through the synchronous belt. The idler pulley is fixed on the idler pulley fixing bracket through the idler pulley shaft. The linear guide is fixed on the guide pad and evenly distributed on both sides of the synchronous belt. The guide rail fixing block is fixed on the fixed base plate. The slider is slidably connected to the linear guide.

[0024] As a further limitation of the second aspect of the invention, the frame assembly also includes a sensor rail, which is fixed to one side of the rail pad, and at least two photoelectric switches are detachably fixed to the sensor rail for detecting the position of the lower coating component. The idler gear fixing bracket can slide up and down along the fixed base plate. The adjusting bolt fixing plate is fixed to the fixed base plate. By tightening or loosening the adjusting bolt on the adjusting bolt fixing plate, the distance between the idler gear fixing bracket and the adjusting bolt fixing plate can be adjusted, thereby tensioning or loosening the timing belt.

[0025] Compared with the prior art, the beneficial effects of the present invention are: This invention innovatively proposes a multi-agent collaborative anti-icing coating control method for overhead cables, and develops an anti-icing coating robot integrating pressing, cleaning, and coating functions. It designs a pressing discrimination agent, a coating effect recognition agent, and a closed-loop decision-making agent, employing a phased operation and agent collaborative feedback scheme. First, the robot's online pressing status is autonomously judged to ensure stable online operation. Then, the robot's coating effect dynamic feedback recognition step enables dynamic adjustment of cleaning and coating. Finally, the robot's fully autonomous lifting and unloading control step completes the operation. The entire process relies on the agent's analysis of image information to achieve key state judgments and strategy optimization. This solves the problems of traditional overhead cable anti-icing coating relying on manual labor, low automation, and the disconnect between state monitoring and strategy adjustment during the coating process. It improves the automation level and efficiency of anti-icing coating operations, enhances the precision control and reliability of the coating process, avoids the safety risks of manual high-altitude operations, and enhances the equipment's adaptability to overhead cable operation scenarios.

[0026] This invention innovatively proposes an autonomous and accurate method for recognizing the clamping state of an overhead cable anti-icing coating robot based on image detection confidence. It designs a complete discrimination process based on image preprocessing, feature extraction, target detection, and comprehensive scoring. Employing a technical solution centered on the YOLOv5 network model, the method first preprocesses the image after the clamping action, extracting enhanced feature maps through the model's CSPDarknet-53 backbone network. Then, the detection head identifies the clamping wheel and cable targets. After screening valid targets, the contact overlap and the mean of the detection confidence are calculated. A comprehensive score is obtained through weighted fusion to determine the clamping state. This method solves the problems of traditional clamping state judgment relying on manual observation or simple sensors, resulting in low discrimination accuracy, and high error rates and interference in target recognition under complex scenarios. It improves the accuracy and stability of clamping state judgment, avoids safety risks such as robot instability and detachment during operation due to clamping state judgment errors, enhances the effectiveness of image feature extraction and the reliability of target detection, and strengthens the adaptability of the discrimination process to complex operating environments.

[0027] This invention innovatively proposes a multimodal dynamic feedback recognition method for the anti-icing coating effect of overhead cables. It designs a collaborative control mechanism for coating strategy, feeding speed, and exposure time, employing a dynamic feedback and environmental adaptation scheme. The robot's walking speed is adjusted based on the coating effect recognition results. The real-time coating feeding speed is determined by combining the actual cable diameter, walking speed, and target coating thickness. Simultaneously, the exposure time for image acquisition is dynamically compensated for changes in ambient temperature, forming a closed-loop control with multi-parameter linkage. This solves the problems of uneven coating thickness caused by mismatch between walking speed and feeding speed in traditional coating processes, as well as the impact of ambient temperature changes on image acquisition quality, thus interfering with effect judgment. It improves the uniformity and consistency of coating thickness, avoids cost waste and insufficient anti-icing performance caused by excessively thick or thin coatings, enhances coating utilization and operational adaptability, and improves the stability and recognition accuracy of image acquisition under different ambient temperatures.

[0028] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0029] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0030] Figure 1 A schematic diagram of the coating module provided as an exemplary embodiment of the present invention Figure 1 ; Figure 2 A schematic diagram of the coating module provided as an exemplary embodiment of the present invention Figure 2 ; Figure 3 A schematic diagram of the upper coating component structure is provided for an exemplary embodiment of the present invention; Figure 4 A schematic diagram of the undercoating component structure provided in an exemplary embodiment of the present invention; Figure 5 A schematic diagram of a floating wheel assembly structure is provided for an exemplary embodiment of the present invention; Figure 6 A schematic diagram of the framework component structure provided for an exemplary embodiment of the present invention; Figure 7 A schematic diagram of an emergency opening component provided as an exemplary embodiment of the present invention; Figure 8 A schematic diagram of a coating control process is provided as an exemplary embodiment of the present invention; Figure 9 A schematic diagram of a triple limit control method provided as an exemplary embodiment of the present invention; Figure 10 A schematic diagram of the communication connection of the control components is provided for an exemplary embodiment of the present invention; Figure 11 A schematic flowchart of an emergency disconnection method for anti-icing coating of overhead power lines, provided as an exemplary embodiment of the present invention; Figure 12 A schematic diagram of the structure of an anti-icing coating robot provided as an exemplary embodiment of the present invention; Figure 13 A flowchart illustrating an exemplary embodiment of the anti-icing coating robot control method provided by the present invention; Figure 14 A schematic diagram of the pressing and identification process of the anti-icing coating robot control method provided as an exemplary embodiment of the present invention; Figure 15 A schematic diagram of the coating recognition process of an anti-icing coating robot control method provided as an exemplary embodiment of the present invention; in, 1-1 Control assembly; 1-2 Upper coating assembly; 1-3 Lower coating assembly; 1-4 Floating wheel assembly; 1-5 Frame assembly; 1-6 Emergency opening assembly; 2-1. Foam cover plate; 2-2. Foam; 2-3. Quick-connect connector; 2-4. Upper coating bracket; 3-1. Sensor detection plate; 3-2. Belt pressure block; 3-3. Slider fixing plate; 3-4. Quick-change connector; 3-5. Lower coating bracket; 3-6. Foam cover plate; 3-7. Second foam; 4-1 Locking pin; 4-2 Driver mounting bracket; 4-3 Floating wheel; 4-4 Left connecting rod; 4-5 Right connecting rod; 4-6 Driver; 4-7 Floating wheel shaft; 5-1. Motor mounting plate; 5-2. Linear guide rail; 5-3. Synchronous belt; 5-4. Idler pulley; 5-5. Idler pulley shaft; 5-6. Adjusting bolt mounting plate; 5-7. Adjusting bolt; 5-8. Synchronous belt pulley; 5-9. Photoelectric switch; 5-10. Drive motor; 5-11. Guide rail pad; 5-12. Sensor guide rail; 5-13. Idler pulley mounting bracket; 5-14. Mounting base plate; 6-1. Belt clamp; 6-2. Emergency pull ring; 7-1 Coating module; 7-2 Coating image acquisition unit; 7-3 Opening and closing motor; 7-4 Upper photoelectric switch; 7-5 Lower photoelectric switch; 7-6 Upper and lower line hoisting mechanism; 7-7 Walking motor; 7-8 Cleaning module; 7-9 Cleaning image acquisition unit; 7-10 Pressing image acquisition unit; 7-11 Pressing module; 7-12 Upper position detection switch; 7-13 Lower position detection switch; 7-14 Control terminal; 7-15 Coating supply system; 7-16 Pressing motor; 7-17 Pressing roller. Detailed Implementation

[0031] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0032] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0033] Overhead power lines are widely used as the primary medium for long-distance power transmission. However, with the northward shift of precipitation lines, these lines experience varying degrees of icing in winter, seriously jeopardizing power grid safety and causing phenomena such as conductor galloping, and in severe cases, even tower collapses and line breaks. Active de-icing methods are numerous and vary significantly in cost; passive coating appears to be more economical and practical. Conventional coating methods involve manual climbing of overhead lines to apply the coating, which compromises safety and is inefficient. Therefore, the development of anti-icing coating robots is urgently needed. In view of this, this implementation proposes a coating module, such as... Figure 1 and Figure 2 As shown, it includes: an upper coating component 1-2, a lower coating component 1-3, an emergency opening component 1-6, a floating wheel component 1-4, and a frame component 1-5. The lower coating component 1-3 is connected to the synchronous belt 5-3 on the frame component 1-5 and can move up and down under the drive of the synchronous belt 5-3 to cooperate with the upper coating component 1-2 to cover the overhead cable.

[0034] like Figure 3 As shown, the upper coating assembly 1-2 includes a first foam cover plate 2-1, a first foam 2-2, a first quick-connect connector 2-3, and an upper coating bracket 2-4. The first foam cover plate 2-1 is fixed on the upper coating bracket 2-4 to form the upper inner cavity of the upper coating assembly 1-2. The first foam 2-2 is fixed in the upper inner cavity and is used to evenly apply the coating to the upper half of the cable. One end of the first quick-connect connector is connected to the coating supply system, and the other end of the first quick-connect connector is fixed on the upper coating bracket 2-4 and communicates with the upper inner cavity.

[0035] like Figure 4As shown, the lower coating assembly 1-3 includes a sensor detection piece 3-1, a first belt pressure block 3-2, a slider fixing plate 3-3, a second quick-connect connector 3-4, a lower coating bracket 3-5, a second foam cover plate 3-6, and a second foam 3-7. The second foam cover plate 3-6 is fixed to the lower coating bracket 3-5, forming the lower inner cavity of the lower coating assembly 1-3. The second foam 3-7 is fixed inside the lower inner cavity and is used to evenly apply the coating to the lower half of the cable. One end of the second quick connector is fixed to the lower coating bracket 3-5 and communicates with the lower cavity. The other end of the second quick connector is connected to the coating recovery system. The timing belt 5-3 is fixed to the first belt pressure block 3-2 on one side of the slider fixing plate 3-3 by bolts. The lower coating bracket 3-5 is fixed on the other side of the slider fixing plate 3-3, and a slider is fixed on the slider fixing plate 3-3. The sensor detection piece 3-1 is fixed to one side of the slider fixing plate 3-3 and is used to detect the position of the lower coating assembly 1-3.

[0036] like Figure 5 As shown, the floating wheel assembly 1-4 includes a locking pin 4-1, a driver mounting bracket 4-2, a floating wheel 4-3, a left connecting rod 4-4, a right connecting rod 4-5, a driver 4-6, and a floating wheel shaft 4-7. The driver 4-6 is fixed to the driver mounting bracket 4-2 and is used to process motor parameters and sensor signals. The left connecting rod 4-4 and the right connecting rod 4-5 are bolted to both sides of the driver mounting bracket 4-2 to form a support frame. The floating wheel shaft 4-7 fixes the floating wheel 4-3 to the bottom of the driver mounting bracket 4-2, supporting the coating mechanism and cable alignment. For cables of different diameters, by replacing the floating wheels 4-3 with different diameters and the inner core diameters of the upper and lower coating assemblies 1-3, it can adapt to various cable diameters and perform coating operations. The driver mounting bracket 4-2 is connected to the robot body of the anti-icing coating robot via the locking pin 4-1 and can slide up and down the robot body.

[0037] like Figure 6 As shown, the frame assembly 1-5 includes a motor mounting plate 5-1, a linear guide rail 5-2, a synchronous belt 5-3, an idler pulley 5-4, an idler pulley shaft 5-5, an adjusting bolt mounting plate 5-6, adjusting bolts 5-7, a synchronous pulley 5-8, a photoelectric switch 5-9, a drive motor 5-10, a guide rail pad 5-11, a sensor guide rail 5-12, an idler pulley mounting bracket 5-13, and a fixed base plate 5-14. The drive motor 5-10 is fixed to the motor mounting plate 5-1 with bolts. The motor mounting plate 5-1 is fixed to the fixed base plate 5-14. The synchronous pulley 5-8 is fixed to the output shaft of the drive motor 5-10 and connected to the idler pulley 5-4 via the synchronous belt 5-3. The idler pulley 5-4 is fixed to the idler pulley mounting bracket 5-13 via the idler pulley shaft 5-5. The idler pulley mounting bracket 5-13 can slide up and down along the fixed base plate 5-14.

[0038] Adjusting bolt fixing plate 5-6 is fixed to fixed base plate 5-14. By tightening or loosening adjusting bolt 5-7, the distance between idler wheel fixing bracket 5-13 and adjusting bolt fixing plate 5-6 is adjusted, thereby achieving tensioning of synchronous belt 5-3. Linear guide rails 5-2 are evenly distributed on both sides of synchronous belt 5-3. They are made to be on the same plane as synchronous belt 5-3 by guide rail pads 5-11 (guide rail pads 5-11 are fixed to fixed base plate 5-14). Sensor guide rail 5-12 is fixed to one side of guide rail pad 5-11 by bolts. At the same time, photoelectric switch 5-9 is fixed to sensor guide rail 5-12 for detecting the position of lower coating component 1-3.

[0039] like Figure 7 As shown, the emergency opening component 1-6 includes a second belt pressure block 6-1 and an emergency pull ring 6-2 (i.e., a lifting device, or a hook). The emergency pull ring 6-2 fixes the synchronous belt 5-3 to the second belt pressure block 6-1 with bolts. The drone lifts the emergency lifting device on the robot through the insulating rod suspended at the bottom, and pulls the emergency pull ring 6-2 upward through the wire rope, which drives the synchronous belt 5-3 to rotate clockwise, causing the lower coating component 1-3 to move downward accordingly, thus completing the opening of the coating module and realizing emergency disconnection.

[0040] The coating module in this implementation can control the upper and lower limits of the lower coating components 1-3, achieving closed-loop control. For example... Figure 8 As shown, the system first initializes and reads the original configuration parameters, then waits to receive control commands from drive motors 5-10; next, it checks whether an opening or closing command has been received. If not, it continues to wait for control commands from drive motors 5-10; if received, it executes the opening or closing operation of drive motors 5-10, then checks whether drive motors 5-10 are in position. If not, it continues to check; if in position, it executes the stopping operation of drive motors 5-10, and then returns to the step of waiting for control commands from drive motors 5-10.

[0041] Specifically, when control component 1-1 receives a closing command, drive motor 5-10 drives synchronous pulley 5-8 to rotate, causing synchronous belt 5-3 to rotate counterclockwise. The lower coating component 1-3, mounted on synchronous belt 5-3, moves upwards. Figure 9 As shown, triple limit control logic is executed: The first layer of logic involves the motor starting from its initial position and automatically stopping after counting the number of pulses corresponding to the position between the upper and lower limits. Drivers 4-6 then feed back the position status to the terminal interface (e.g., ...). Figure 9 (including upper and lower flow limit protections) to facilitate the next step of the operation; The second layer of logic involves the coating component moving upwards, triggering the upper limit photoelectric sensor (e.g., the photoelectric detection sheet on it) to activate the upper limit photoelectric sensor. Figure 9(The upper and lower photoelectric switches in the middle) After the driver 4-6 collects the signal, the driver 4-6 controls the drive motor 5-10 to stop and feeds back the sensor status to the terminal interface for the next step of operation; The third layer of logic involves the coating component moving upwards and contacting a mechanical limit plate (e.g., the upper end of a linear guide 5-2), causing the motor to stall. The control board collects this stall information, stops the motor, and feeds back the stall status to the terminal interface for further processing. For example... Figure 9 In this case, a time-delay protection method can be used for limit control.

[0042] Conversely, when the lower coating components 1-3 are released, the same triple logic control is used. The anti-icing coating robot achieves limit protection closed-loop control through this triple logic coupling control, making the position of the lower coating components 1-3 controllable.

[0043] Figure 10 The communication control logic of control component 1-1 is also shown. Control component 1-1 is connected to the control terminal of the anti-icing coating robot via CAN bus. Control component 1-1 is powered by battery. Control component 1-1 is connected to drive motor 5-10 and photoelectric switch 5-9 (including upper photoelectric switch 7-4 and lower photoelectric switch 7-5).

[0044] In summary, this invention provides a coating module adaptable to various wire diameters to address complex and variable on-site conditions, thereby improving the robot's applicability. It proposes an emergency opening strategy using the 5-3 synchronous belt's opposite-side reversal principle, which allows the coating mechanism to be opened in reverse to release the robot from its locked state. A coating up-and-down motion structure is constructed, employing a 5-3 synchronous belt structure and a 5-2 linear guide rail, improving the stability of the coating operation. Furthermore, by integrating multiple sensors, a multi-coupled, highly reliable multi-protection logic for the coating mechanism's limit positions is built, overcoming the shortcomings of existing limit methods such as poor stability and susceptibility to malfunctions, thus improving operational safety and stability, and enhancing the robot's coating effect.

[0045] Based on the above coating module, Figure 11 An emergency disconnection method for anti-icing coating on overhead power lines is shown, comprising the following steps: S1101: The current coating component cannot detach from the overhead cable, or the lower coating component, although detached from the overhead cable, cannot move to the set disconnection position, and is determined to be in an emergency disconnection state. S1102: In an emergency disconnection situation, an emergency lifting device is pulled by a drone, which drives the synchronous belt to move the lower coating component away from the upper coating component; S1103: Emergency de-coring is completed when the current coating component moves to the set de-coring position.

[0046] In S1102 of this implementation, depending on the different control subjects, multiple control methods may be included, specifically including: Method 1: In emergency disconnection state, the controller of the anti-icing coating device (or the controller shared with the anti-icing coating) automatically sends an emergency disconnection request to the drone, so that the drone can pull the emergency lifting device through the suspended boom according to the received request; Method 2: In the emergency disconnection state, the controller of the anti-icing coating device (or the controller shared with the anti-icing coating) sends an emergency disconnection request to the ground control terminal, so that the UAV can control the UAV to pull the emergency lifting equipment according to the received request; Method 3: In the event of an emergency disconnection, the emergency lifting equipment can be pulled by a drone directly through a ground control terminal.

[0047] This implementation proposes an anti-icing coating robot, such as Figure 12 As shown, it includes: a robot body and cleaning module 7-8, coating module 7-1, clamping module 7-11, control terminal 7-14, walking wheel set, clamping image acquisition unit 7-10 (realized by a camera) and coating image acquisition unit 7-2 (realized by a camera) arranged on the robot body. The cleaning module 7-8 is located at the front end of the robot body, the coating module 7-1 is located at the rear end of the robot body, the clamping module 7-11 is located in the middle of the robot body, and the walking wheel set includes a first wheel set and a second wheel set. The first wheel set is located between the cleaning module 7-8 and the clamping module 7-11, and the second wheel set is located between the coating module 7-1 and the clamping module 7-11. The first wheel set and the second wheel set are driven by the walking motor 7-7 on the robot body. The control terminal 7-14 is communicatively connected to the cleaning module 7-8, the coating module 7-1, the pressing module 7-11, the control terminal 7-14, the walking wheel set, the pressing image acquisition unit 7-10, and the coating image acquisition unit 7-2.

[0048] The clamping module 7-11 includes: a clamping roller 7-17, a clamping motor 7-16, an upper position detection switch 7-12 (i.e., a first position switch), and a lower position detection switch 7-13 (i.e., a second position switch). The clamping roller 7-17 can move up and down under the drive of the clamping motor 7-16 to clamp or loosen the cable. The upper position detection switch 7-12 and the lower position detection switch 7-13 are arranged vertically. The upper position detection switch 7-12 is used to identify the clamping state, and the lower position detection switch 7-13 is used to identify the loosening state.

[0049] The coating module 7-1 includes a coating opening and closing module, an opening and closing motor 7-3, an upper photoelectric switch 7-4 (i.e., a first limit switch), and a lower photoelectric switch 7-5 (i.e., a second limit switch). The coating opening and closing module can move up and down under the drive of the opening and closing motor 7-3 to approach or move away from the cable to be coated for coating (the nozzle is arranged on the inner wall of the two segments of the coating opening and closing module). The first limit switch and the second limit switch are arranged vertically. The first limit switch is used to limit the upper limit position of the coating opening and closing module, and the second limit switch is used to limit the lower limit position of the coating opening and closing module.

[0050] The upper part of the robot body is connected to a lifting mechanism 7-6 for use in conjunction with a drone to achieve loading and unloading.

[0051] Based on the aforementioned anti-icing coating robot, such as Figure 13 As shown, this implementation proposes a control method for an anti-icing coating robot, including the following process: The S1301 robot's autonomous judgment process for clamping status is as follows: After the anti-icing coating robot is hoisted onto the cable, the clamping module is controlled to perform the clamping action. The clamping judgment agent is called, and a clamping judgment result is generated based on the image detection confidence after the clamping action is completed. When the clamping judgment result is that the robot is clamped, the anti-icing coating robot is online. S1302 robot coating effect dynamic feedback recognition steps: start the cleaning module to clean the cable, control the anti-icing coating robot to walk along the cable, control the coating module to coat the cleaned cable, call the coating effect recognition intelligent agent, generate the coating effect recognition result based on the image after coating, call the closed-loop decision intelligent agent to dynamically adjust the coating strategy of the coating module based on the coating effect recognition result. S1303 Robot Fully Autonomous Lifting and Unloading Control Steps: After the anti-icing coating robot completes the coating action, the clamping module is controlled to perform the loosening action. The clamping judgment agent is called to generate a loosening judgment result based on the image after the loosening action is completed. When the loosening judgment result is that it has been loosened, the anti-icing coating robot lifting and unloading process is executed.

[0052] Steps S1301 and S1303 of this implementation are the processes of tightening the upper wire and loosening the lower wire, respectively. Specifically, they include the following processes: The clamping module 7-11 is an important component of the anti-icing coating robot. Its main function is to ensure that the anti-icing coating robot is stably attached to the cable during the anti-icing coating operation, and to maintain sufficient gripping force even in harsh weather conditions, so as to prevent the anti-icing coating robot from falling off the cable, thus ensuring the safety and efficiency of the operation. Through precise mechanical design and electronic control, the clamping module 7-11 realizes flexible opening and closing adjustment of the clamping structure, which can adapt to overhead cables of different diameters and types, and ensure that the anti-icing coating robot is firmly clamped on the cable.

[0053] More specifically, the clamping recognition process can include five steps: control signal input, clamping motor 7-16 action, position detection switch input, image processing result output, and status feedback output. For example... Figure 14 As shown, after the clamping recognition begins, a control signal is input first, followed by two processes. One process is to perform a clamping operation, causing the clamping motor 7-16 to rotate forward. When the upper limit is triggered, the camera (i.e., the clamping image recognition unit 7-10) outputs information and determines whether the clamping is complete. If the clamping is not complete, the clamping motor 7-16 continues to rotate forward; if the clamping is complete, the clamping status is fed back, and the robot is finally ready to go offline. The other process is to perform a releasing operation, causing the clamping motor 7-16 to rotate in reverse. When the lower limit is triggered, the camera outputs information and determines whether the robot is released. If the robot is released, the release status is fed back, and the robot is ready to go offline; if the robot is not released, the clamping motor 7-16 continues to rotate in reverse.

[0054] More specifically, it includes the following: (1) Control signal input: After the UAV hoists the anti-icing coating robot onto the cable or before the anti-icing coating robot is taken off the line, it receives clamping control instructions from the control terminal 7-14, including trigger signals for clamping and releasing actions; (2) Operation of clamping motor 7-16: After receiving the clamping signal, clamping motor 7-16 starts to move forward, and clamping wheel 7-17 moves upward to clamp the cable; after receiving the release signal, clamping motor 7-16 starts to move in reverse, and clamping wheel 7-17 moves downward to release the cable. (3) Position detection switch input: Position detection switches are installed at the upper and lower ends of the travel of the clamping module 7-11. The signals reflecting the opening and closing degree of the clamping mechanism are received from the position detection switches and used for real-time monitoring and control of the clamping force. During the movement of the clamping motor 7-16, when the baffle of the clamping module 7-11 touches the upper position detection switch 7-12 (i.e., the first position detection switch), it indicates that the clamping module 7-11 has been clamped; conversely, when the lower position detection switch 7-13 (i.e., the second position detection switch) is triggered, it indicates that the clamping module 7-11 has been released.

[0055] (4) Output of camera image processing results: After the clamping module 7-11 is clamped or loosened, the clamping image acquisition unit 7-10 installed above the clamping module 7-11 of the anti-icing coating robot takes pictures of the position of the clamping module 7-11 and the cable and uploads them to the clamping discrimination agent of the control terminal 7-14, and outputs the image processing results. The image processing process is as follows: 4-1) Model training phase: First, the collected images of compression and release are classified and labeled, and the image samples are divided into training set and test set according to the proportion; then, the deep learning model with improved YOLOv5 network architecture is fine-tuned and trained; finally, the trained model is used to test the test set and evaluate the model to obtain the image processing model Model-0 (i.e. the basic framework of the compression discrimination agent). 4-2) Image judgment stage of pressing in place: After the shooting is completed, the image data in the camera is transmitted to a dedicated control terminal 7-14 (equipped with a pressing discrimination agent, or it can be a control terminal on a ground remote controller or a ground mobile terminal). The pressing discrimination agent automatically analyzes and processes the captured image, uses the trained model Model-0 to judge the uploaded image, determines whether the pressing module 7-11 is in place, and outputs the judgment result.

[0056] Optionally, the improved YOLOv5 network architecture of this invention can adopt a commercially available improved YOLOv5 network architecture (as the basic framework for the compression discriminative agent); or, the compression and loosening determination can be performed in the following ways, specifically including: Step 1: Input the RGB image after the clamping action is performed. Preprocessing is used to eliminate noise caused by uneven outdoor lighting and slight shaking during hoisting, and a clear image is output to lay the foundation for subsequent feature extraction.

[0057] Step 2: Based on the preprocessed clear image, extract global basic features using the native YOLOv5 backbone network, add a lightweight edge enhancement layer to highlight the edge and contour features of the "pressure wheel-cable", and output the enhanced feature map. This improves the accuracy of target detection.

[0058] ; in, The original feature map representing the output of the CSPDarknet-53 backbone network of YOLOv5; Represents the Sobel edge detection operator; Represents the edge feature enhancement coefficient; Step 3: YOLOv5 detection head based on enhanced feature maps It accurately detects two core targets: "pressure rollers" and "cables," outputting the target's location, size, confidence level, and category. Valid targets are obtained through confidence level screening, providing reliable objects for subsequent contact state quantification.

[0059] ; in, A complete set of detection results representing a single target; The horizontal center coordinates of the target bounding box; Represents the vertical center coordinates of the target bounding box; Represents the width of the target bounding box; Represents the height of the target bounding box; This represents the confidence level of the target detection.

[0060] when At that time, it is an effective target, thus obtaining the effective pressure wheel target. and effective cable target ,in, The target detection confidence threshold.

[0061] Step 4: Based on effective objectives and The system calculates a comprehensive score by fusing two indicators: "Interaction Overlap (IoU) + Mean Detection Confidence". Then, it uses a fixed threshold to decide whether to "press or release", and outputs the final recognition result.

[0062] Contact overlap for: ; in, This represents the function for calculating the area of ​​the bounding box. This represents the intersection area between the effective clamping wheel and the effective cable boundary frame; This represents the union region of the effective clamping wheel and the effective cable boundary frame; The area represents the intersection region; Represents the area of ​​the union region.

[0063] The mean confidence level of the detection is: ; in, and These represent the effective clamping wheel targets. and effective cable target The confidence level.

[0064] To detect the mean confidence level With contact overlap The weighted sum is used as the comprehensive score. ,include: ; in, The weight representing the degree of contact overlap. The weights represent the mean confidence level of the detection.

[0065] The final decision was: ; Where T is the set threshold.

[0066] (5) Output the status feedback and fine-tune the clamping module 7-11. Call the closed-loop decision-making agent to fine-tune the clamping module 7-11. There are four possible scenarios: Scenario 1: The control signal input is a clamping signal, the upper position detection switch 7-12 is triggered, the image judgment result is clamping, the status feedback output is clamping, and the mobile terminal displays that it has been clamped, indicating that the anti-icing coating robot can move. Scenario 2: The control signal input is a clamping signal, and the mobile terminal displays that it is clamping. At this time, the control terminal sends a motor forward rotation command until the image judgment result is clamping, the status feedback output is clamping, and the mobile terminal communicating with the control terminal 7-14 displays that it is clamping, and the anti-icing coating robot can move. Scenario 3: The control signal input is a release signal, the lower position detection switch 7-13 is triggered, the image judgment result is release, the status feedback output is release, and the mobile terminal displays that it has been released. The anti-icing coating robot can be taken offline. Scenario 4: When the control signal input is a release signal, the lower position detection switch is triggered, the image judgment result is "pressed", the status feedback output is "releasing", and the mobile terminal displays "releasing". At this time, the control terminal sends a motor reverse command until the image judgment result is "releasing", the status feedback output is "releasing", and the mobile terminal displays "releasing". The anti-icing coating robot can then be taken offline.

[0067] Step S202 of this implementation is the cleaning and coating process, such as... Figure 15 As shown, after the coating recognition begins, a control signal is input first, followed by the cleaning module, then the paint supply, and then the walking motor. After that, the camera outputs information and determines whether the coating is uniform. If it is uniform, it continues to work at this flow rate. If it is uneven, it is divided into two cases: too much paint and too little paint. If there is too much paint, the flow rate is reduced, and if there is too little paint, the flow rate is increased. After that, it returns to the camera output stage.

[0068] In this implementation, when the pressing status feedback is "pressed" and the pressing is displayed on the mobile terminal, it indicates that the anti-icing coating robot has been successfully put into operation. At this time, the paint supply system 7-15 is activated, and the anti-icing paint carried by the anti-icing coating robot is delivered to the coating component through the liquid supply system to perform anti-icing coating operations on the line. Coating image acquisition units 7-2 are installed in front of and behind the anti-icing coating robot. These units communicate with the ground-based remote controller via a wireless receiving module, acquiring image information in real time during the coating process. This information is displayed on the ground-based remote controller, which can monitor the anti-icing coating robot's line status in real time and adjust the paint supply speed in real time based on the coating effect to ensure uniform coating of the line. Alternatively, the images can be directly transmitted to the anti-icing coating robot's control terminal 7-14 for adaptive control.

[0069] The coating effect recognition process is as follows: (1) Control signal input: After the anti-icing coating robot is online, it receives the cable cleaning signal from the control terminal 7-14 of the anti-icing coating robot. The cleaning module 7-8 installed in front of the anti-icing coating robot starts to work. At this time, the cleaning motor drives the cleaning brush head to clean the cable (the cleaning image acquisition unit 7-9 (preferably a camera) acquires the image, and the cleaning effect is identified by the cleaning recognition intelligent agent. If it is not qualified, the anti-icing coating robot stops walking and continues cleaning until the cleaning effect is qualified). (2) The paint supply system 7-15 works: The paint supply speed is matched with the initial feeding speed according to the cable specifications. After the cleaning work starts, the coating operation control signal is received from the control terminal 7-14 of the anti-icing coating robot. The paint supply system 7-15 starts to supply paint to the coating module 7-1 installed behind the anti-icing coating robot according to the initial feeding speed to perform coating operation on the cable. (3) Walking motor 7-7 action: At the same time, the control terminal 7-14 of the anti-icing coating robot receives the forward or backward command issued by the mobile terminal to the anti-icing coating robot, and the walking motor 7-7 starts to move. During the walking process, the paint supply system 7-15 delivers paint to the coating module 7-1 to spray the cable. (4) Image processing results output of coating image acquisition unit 7-2: The coating image acquisition unit 7-2, installed on the anti-icing coating robot body, takes pictures during the anti-icing coating process and uploads them to the control terminal 7-14 (which can also be placed on a remote control or mobile terminal), and outputs the image processing results. The image processing process is as follows: 4-1) Model Training Phase: First, images of uncoated cables collected from the front of the anti-icing coating robot are compared with images of coated cables collected from the rear of the robot. The images are categorized and labeled as samples with insufficient coating, uniform coating, and excessive coating. The image samples are then divided into training and test sets according to the proportions. Next, a deep learning model with an improved YOLOv5 network architecture is used for fine-tuning and training. Finally, the trained model is used to test and evaluate the test set to obtain the image processing model Model-0 (i.e., the basic framework of the coating effect recognition agent). 4-2) Coating effect image judgment stage: After the image is captured, the image data in the coating image acquisition unit 7-2 is transmitted to the dedicated control terminal 7-14 (equipped with coating effect recognition intelligent agent). The coating effect recognition intelligent agent automatically analyzes and processes the captured image, judges the uploaded image, determines whether the coating effect is uniform, and outputs the judgment result.

[0070] (5) The closed-loop decision-making agent adjusts the paint supply speed in real time, and fine-tunes the paint supply speed based on the state feedback output. There are three cases: Scenario 1: If the image judgment result indicates insufficient coating, increase the paint supply speed until the image judgment result indicates uniform coating. At this time, the anti-icing coating robot will supply material at this speed. Scenario 2: If the image determination result indicates uniform spraying, maintain this feeding speed. In this case, the anti-icing coating robot will feed materials at this speed. Scenario 3: If the image judgment result indicates excessive spraying, reduce the paint supply speed until the image judgment result indicates uniform spraying. At this time, the anti-icing coating robot will supply material at this speed.

[0071] More specifically, in real-world applications, when the walking speed increases or the cable diameter increases, the coating may become too thin; conversely, the coating may become too thick, leading to paint waste or insufficient anti-icing effect. Therefore, in this implementation, the closed-loop decision-making agent determines the real-time paint supply speed based on the actual cable diameter, the anti-icing coating robot's walking speed, and the target coating thickness, including: ; in, This represents the coating compatibility correction factor. Represents pi (π) Represents the actual outer diameter of the cable. This represents the walking speed of the anti-icing coated robot. Represents the target coating thickness. Represents the real-time feed rate of the coating, with the dimension in meters (m). 3 / s.

[0072] This dynamic adjustment strategy for the real-time coating feeding speed achieves dynamic matching between the feeding speed and cable diameter, travel speed, and target thickness, avoiding coating thickness deviation while ensuring the effective thickness of the anti-icing coating.

[0073] Similarly, in real-world applications, when the viscosity of the coating suddenly increases (e.g., in low-temperature environments), the coating thickness will increase even if the feeding speed remains constant. In such cases, the walking speed needs to be adjusted synchronously to maintain uniformity; otherwise, the feeding system requires frequent adjustments, resulting in lag. Therefore, the closed-loop decision-making agent dynamically adjusts the walking speed of the anti-icing coating robot in real time, including: ; in, This represents the corrected walking speed; Represents the initial walking speed; Represents the target coating thickness; This represents the actual coating thickness. Adjusting the auxiliary feeding system by modifying the travel speed reduces frequent operation of the feeding valve (extending valve life) and shortens the response time for coating thickness adjustment. For example, if... > (If the coating is too thick), then > Increase walking speed to reduce the amount of coating adhering to a unit length of cable; if < (If the coating is too thin), then Slow down walking speed to increase paint adhesion.

[0074] It should be noted that the image acquisition units in this implementation all use low-temperature cameras. However, low temperatures (such as -10℃ to -25℃) will cause a decrease in the sensitivity of the camera's photosensitive element. If a fixed exposure time is maintained, details in dark areas of the image will be lost (such as the contact points between the clamping rollers 7-17 and the cable), leading to increased errors in the judgment of clamping recognition and coating effect recognition. In view of this, this implementation proposes an exposure time compensation strategy, specifically including: ; in, This represents the exposure time after compensation; Represents the baseline exposure time; This represents the temperature compensation coefficient, with dimensions of the reciprocal of temperature. Represents the reference temperature; This represents the actual ambient temperature.

[0075] when Lower (e.g., from 0°C to -20°C). Automatically extend the time (e.g., from 20ms to 28ms) to avoid overly dark images; when rise, Shorten the image to avoid overexposure.

[0076] To further improve the coating effect, this implementation also proposes a coating thickness prediction scheme, including: ; in, Represents the predicted coating thickness; Represents the real-time feeding speed; This represents the solids content of the coating. Represents pi; This represents the actual outer diameter of the cable; Represents real-time walking speed.

[0077] The above solution enables pre-prediction of coating thickness, avoiding rework and reducing paint waste. and In contrast, when When the current parameters are deemed suitable, the operation continues; when... At that time, decrease Q or increase it in advance. Avoid making it too thick; when Increase or decrease Q in advance Avoid making it too thin.

[0078] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An overhead cable multi-agent collaborative anti-icing coating control method, characterized in that, An ice prevention coating robot comprising a compacting module, a cleaning module, and a coating module, comprising the following processes: A robot online compacting state autonomous judgment step: after the ice prevention coating robot is hoisted onto the cable, the compacting module is controlled to perform a compacting action, a compacting discrimination intelligent agent is called, a compacting judgment result is generated according to the image detection confidence after the compacting action is completed, and when the compacting judgment result is compacted, the ice prevention coating robot is completed online; A robot coating effect dynamic feedback identification step: the cleaning module is started to clean the cable, the coating module is controlled to coat the cleaned cable, a coating effect identification intelligent agent is called, a coating effect identification result is generated according to the image after coating is completed, and a closed-loop decision intelligent agent is called to dynamically feedback and adjust the coating strategy of the coating module according to the coating effect identification result; A robot fully autonomous hoisting and offline control step: after coating is completed, the compacting module is controlled to perform a loosening action, a compacting discrimination intelligent agent is called, a loosening judgment result is generated according to the image after the loosening action is completed, and when the loosening judgment result is loosened, the ice prevention coating robot hoisting and offline process is executed.

2. The overhead cable multi-agent collaborative ice prevention coating control method of claim 1, wherein In the robot online compacting state autonomous judgment step, a compacting discrimination intelligent agent is called, a compacting judgment result is generated according to the image detection confidence after the compacting action is completed, comprising: The image after the compacting action is completed is preprocessed, and enhanced feature maps are extracted from the CSPDarknet-53 backbone network of the YOLOv5 network model according to the preprocessed image; Take the enhanced feature map as the input of the detection head of the YOLOv5 network model, detect the compacted wheel and cable targets, output the target bounding box, confidence and category, and screen the effective compacted wheel targets with confidence not lower than a threshold and effective cable targets ​ Computing The contact overlap degree The contact overlap degree The detection confidence mean value The detection confidence mean value The comprehensive score The comprehensive score ; When the comprehensive score If the comprehensive score is greater than or equal to a set threshold, it is determined that the compression is tight; otherwise, it is determined that the compression is loose.

3. The overhead cable multi-agent collaborative ice prevention coating control method of claim 2, wherein Computing With Contact overlap , comprising: ； wherein is an area calculation function.

4. The overhead cable multi-agent collaborative ice prevention coating control method of claim 2, wherein detection confidence mean is: , and represent the confidence of the effective clamp wheel target and the effective cable target respectively; with a detection confidence mean with a contact overlap weighted sum as an overall score .

5. The overhead cable multi-agent collaborative anti-icing coating control method of claim 1, wherein, comprising the following processes: In the robot coating effect dynamic feedback identification step, a closed-loop decision intelligent agent is called to dynamically feedback and adjust the coating strategy of the coating module according to the coating effect identification result, comprising: The actual coating thickness is determined according to the coating effect identification result, and the walking speed of the ice prevention coating robot is dynamically adjusted according to the target coating thickness and the actual coating thickness, comprising: ； wherein, represents the corrected walking speed; represents the initial walking speed; represents the target coating thickness; represents the actual coating thickness.

6. The overhead cable multi-agent collaborative anti-icing coating control method of claim 5, wherein, comprising the following processes: According to the actual diameter of the cable, the walking speed of the anti-icing coating robot and the target coating thickness, the real-time coating supply speed is determined, comprising: wherein, represents a coating adaptation correction coefficient, represents a circle constant, represents an actual outer diameter of the cable, represents a walking speed of the anti-icing coating robot, represents a target coating thickness, represents a real-time coating supply speed.

7. The overhead cable multi-agent collaborative ice prevention coating control method of claim 1, wherein In the image acquisition process after the compacting action is completed, the image acquisition process after coating is completed, and the image acquisition process after the loosening action is completed, dynamic compensation of exposure time is performed, comprising: ； wherein, represents the compensated exposure time; represents the reference exposure time; represents the temperature compensation coefficient; represents the reference temperature; represents the actual ambient temperature.

8. The overhead cable multi-agent collaborative ice prevention coating control method of claim 1, wherein The coating module comprises an upper coating assembly, a lower coating assembly, and a frame assembly, the lower coating assembly is connected with the synchronous belt on the frame assembly and can move up and down under the drive of the synchronous belt to cooperate with the upper coating assembly to coat the overhead cable, and an emergency lifting appliance is connected on the synchronous belt; When the lower coating assembly cannot be separated from the overhead cable, or the lower coating assembly can be separated from the overhead cable but cannot be moved to the set off-line position, it is determined to be in an emergency off-line state; In the emergency off-line state, the unmanned aerial vehicle pulls the emergency sling to drive the synchronous belt to move the lower coating assembly away from the upper coating assembly. The emergency off-line is completed when the lower coating assembly moves to the set off-line position.

9. An anti-icing coating robot characterized by, Comprise: The robot body and the cleaning module, the pressing module, the control terminal, the walking wheel set, the pressing image acquisition unit, the coating image acquisition unit and the coating module arranged on the robot body; The cleaning module is arranged at the front end of the robot body, the coating module is arranged at the rear end of the robot body, the pressing module is arranged at the middle position of the robot body, the walking wheel set comprises a first wheel set and a second wheel set, the first wheel set is arranged between the cleaning module and the pressing module, the second wheel set is arranged between the coating module and the pressing module, and the first wheel set and the second wheel set are driven by the walking motor on the robot body; The control terminal is respectively connected with the cleaning module, the coating module, the pressing module, the control terminal, the walking wheel set, the pressing image acquisition unit and the coating image acquisition unit.

10. The anti-icing coating robot of claim 9, wherein The pressing module comprises a pressing wheel, a pressing motor, a first position switch and a second position switch, the pressing wheel can move up and down under the driving of the pressing motor to press the cable or loosen the cable, the first position switch and the second position switch are arranged up and down, the first position switch is used to identify the pressing state, and the second position switch is used to identify the loosening state.

11. The anti-icing coating robot of claim 9, wherein The coating module comprises an upper coating assembly, a lower coating assembly, a floating wheel assembly and a frame assembly, the lower coating assembly is connected with the synchronous belt on the frame assembly and can move up and down under the driving of the synchronous belt to coat the overhead cable with the upper coating assembly; An emergency sling is connected on the synchronous belt, the emergency sling is used to drive the synchronous belt to move the lower coating assembly away from the upper coating assembly when emergency off-line is needed; The floating wheel assembly comprises a support frame and a floating wheel arranged on the support frame for walking on the overhead cable, the upper coating assembly is connected with the support frame, and the support frame is used to connect with the robot body of the anti-icing coating robot.

12. The anti-icing coating robot of claim 11, wherein The upper coating assembly comprises a first foam cover plate, a first foam, a first quick connector and an upper coating support; The first foam cover plate is fixed on the upper coating support to form an upper inner cavity with an opening downward, the first foam is fixed in the upper inner cavity to apply coating to the overhead cable, one end of the first quick connector is connected with a coating supply system, and the other end of the first quick connector is fixed on the upper coating support and communicates with the upper inner cavity.

13. The anti-icing coating robot of claim 11 or 12, wherein The floating wheel assembly comprises a locking pin, a support frame, a floating wheel and a driver, the support frame comprises a left connecting rod, a right connecting rod and a driver fixing frame connected between the left connecting rod and the right connecting rod, and the driver is fixed on the driver fixing frame; The floating wheel rotating shaft fixes the floating wheel on the bottom of the driver fixing frame, and the support frame is connected with the robot body of the coating robot through the locking pin and can slide up and down along the robot body.

14. The anti-icing coating robot of claim 11, wherein, The lower coating assembly comprises a sensor detection sheet, a belt pressing block, a sliding block fixing plate, a second quick connector, a lower coating support, a second foam, and a second foam cover plate. The second foam cover plate is fixed on the lower coating support to form a lower inner cavity with an opening upward, and the second foam is fixed in the lower inner cavity to coat the coating on the overhead cable. One end of the second quick connector is connected with the coating recovery system, and the other end of the second quick connector is fixed on the lower coating support and communicates with the lower inner cavity. The other side of the sliding block fixing plate is fixedly connected with the lower coating support, the sensor detection sheet is fixed on one side of the sliding block fixing plate to detect the position of the lower coating assembly, and the sliding block is connected on the sliding block fixing plate.

15. The anti-icing coating robot of claim 11, wherein, The frame assembly comprises a motor fixing plate, a linear guide rail, a synchronous belt, an idler, an idler rotating shaft, an adjusting screw fixing plate, an adjusting screw, a synchronous pulley, a photoelectric switch, a driving motor, a guide rail cushion block, an idler fixing frame, and a fixed bottom plate. The driving motor is fixed on the motor fixing plate through a bolt, and the motor fixing plate is fixed on the fixed bottom plate. The synchronous pulley is fixed on the output shaft of the driving motor, the synchronous pulley is connected with the idler through the synchronous belt, the idler is fixed on the idler fixing frame through the idler rotating shaft, the linear guide rail is fixed on the guide rail cushion block and is evenly distributed on both sides of the synchronous belt, the guide rail fixing block is fixed on the fixed bottom plate, and the sliding block is slidably connected with the linear guide rail.

Images