Submarine cable fault detection robot based on bionic inchworm structure

Abstract

The application discloses a submarine cable fault detection robot based on a bionic inchworm structure and relates to the technical field of underwater robots. The robot comprises a first ring-around mechanism, a cleaning mechanism, a telescopic bending mechanism and a second ring-around mechanism which are connected in sequence. The cleaning mechanism integrated in the front end can automatically remove the attachments on the surface of the cable before detection, thereby improving the detection accuracy. The first ring-around mechanism, the telescopic bending mechanism and the second ring-around mechanism construct a bionic inchworm structure, so that the robot can step over inherent obstacles without detaching from the cable. When large obstacles and buried cable sections are crossed, the decoupling propulsion assembly is decoupled to travel and track the cable path, and when the coupling condition is reached, the robot is returned to continue the ring-around crawling detection through the bionic inchworm structure, thereby realizing long-distance, uninterrupted and fully-automatic fault detection on the complex submarine cable. The application improves the quality of submarine cable detection data, the operation efficiency and the intelligent level.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of underwater robots, in particular to a seabed cable fault detection robot based on a bionic inchworm structure. BACKGROUND

[0002] With the continuous development of cross-border communication and offshore wind power, the safe operation and maintenance of underwater cable facilities face great challenges. Traditional seabed cable detection and maintenance mainly rely on manual diving operations, which have the disadvantages of low efficiency, high cost and high risk. Especially in deep sea high pressure, low visibility and other harsh environments, traditional methods are difficult to achieve reliable fault identification and rapid positioning.

[0003] Underwater remotely operated vehicles are gradually applied in seabed cable laying, detection and maintenance, however, the current remotely operated vehicles for detecting seabed cables mainly rely on carrying a separate ordinary optical camera for imaging and manual real-time continuous visual observation and judgment. This kind of remotely operated vehicle hovering near the cable for observation by the optical camera carried has obvious disadvantages. First, underwater currents and swells are easy to cause the detection platform to sway, causing the detection target to be out of the camera's field of view, image blur, and detection data distortion; second, hovering operation has high energy consumption, short endurance time and difficult control; in addition, biological attachment and sediment coverage on the surface of the cable will seriously block the view, leading to missed detection and false detection. On the other hand, traditional attached crawling detection devices cannot continue to travel when encountering cable joints, repair sleeves and other obstacles or when part of the cable is buried, resulting in forced interruption of the detection operation.

[0004] Therefore, there is an urgent need for an intelligent seabed cable fault detection robot that can operate stably, efficiently and continuously, and overcome the above-mentioned adverse conditions. SUMMARY

[0005] To solve the above technical problems, the present application proposes a seabed cable fault detection robot based on a bionic inchworm structure, which can realize long-distance, uninterrupted, full-automatic continuous detection and fault diagnosis operation through C-shaped ring crawling structure, autonomous cleaning and multi-mode obstacle crossing capability.

[0006] To achieve the above purpose, the technical scheme of the present application is as follows:

[0007] The submarine cable fault detection robot based on bionic inchworm structure comprises a first ring mechanism, a cleaning mechanism, a telescopic bending mechanism and a second ring mechanism connected in sequence, wherein the first ring mechanism and the second ring mechanism are both C-shaped and each comprises a fixed arc-shaped arm and a movable arc-shaped arm hinged by a hinge assembly; a rear end anchoring assembly and a front end anchoring assembly are respectively arranged on the first ring mechanism and the second ring mechanism, and each of the front end anchoring assembly and the rear end anchoring assembly comprises a plurality of anchoring foot units; a first driving assembly and a second driving assembly are respectively arranged in the first ring mechanism and the second ring mechanism, and each of the first driving assembly and the second driving assembly comprises a plurality of omnidirectional drive wheel groups; a sensing assembly and a decoupling propulsion assembly are arranged on the outer wall of the first ring mechanism and the second ring mechanism, the sensing assembly comprises a visual camera and a search sonar, and the decoupling propulsion assembly comprises a plurality of decoupling propellers.

[0008] The cleaning mechanism is C-shaped and comprises a first arc-shaped arm and a second arc-shaped arm hinged by a hinge assembly, the first arc-shaped arm and the second arc-shaped arm are both provided with a cleaning assembly, and a strip-shaped sliding hole is formed in the first arc-shaped arm and the second arc-shaped arm, the cleaning assembly comprises a C-shaped cleaning brush, a connecting rod, a sleeve and a cleaning motor, a threaded rod is connected to the output shaft of the cleaning motor through a transmission gear box, one end of the connecting rod is connected to the C-shaped cleaning brush, the other end of the connecting rod passes through the strip-shaped sliding hole and is connected to the sleeve, the sleeve is sleeved on the threaded rod, and the sleeve is threadedly connected with the threaded rod.

[0009] The telescopic bending mechanism comprises a flexible main body, a plurality of C-shaped skeletons and N telescopic motors, the plurality of C-shaped skeletons are embedded in the flexible main body, a plurality of groups of parallel distributed guide holes are arranged on the plurality of C-shaped skeletons, the telescopic motors are arranged on the cleaning mechanism, the telescopic motors are connected with a winch mechanism, and the tendons wound on the winch mechanism are fixedly connected with the second ring mechanism after passing through each group of guide holes.

[0010] Preferably, the hinge assembly comprises a male arm, a female arm, a hinge shaft, a pushing lead screw and a pushing motor, the male arm and the female arm are hinged through the hinge shaft, the pushing motor is arranged on the female arm, the output end of the pushing motor is connected with one end of the pushing lead screw through a shaft connector, and the other end of the pushing lead screw is connected with a hinge ball socket on the male arm through a hinge ball head.

[0011] Preferably, the plurality of anchoring foot units are distributed at equal distances, and each anchoring foot unit comprises a lead screw stepping motor arranged on the outer wall of the fixed arc-shaped arm or the movable arc-shaped arm, the output end of the lead screw stepping motor extends into the fixed arc-shaped arm or the movable arc-shaped arm, and the end of the output end is provided with an anchoring foot.

[0012] Preferably, several omni-directional drive wheel groups are distributed equidistantly; the omni-directional drive wheel group comprises a drive ball head, an elastic support frame, a wheel hub motor, and a drive wheel integrated on an outer rotor of the wheel hub motor, the wheel hub motor is installed on the elastic support frame, and the top of the elastic support frame is connected with a drive ball socket on a fixed or movable arc-shaped arm through the drive ball head.

[0013] Preferably, the wheel surface of the drive wheel is a concave arc surface.

[0014] Preferably, the first arc-shaped arm and the second arc-shaped arm are each provided with a pressurized water spraying assembly on the side wall close to the first surrounding mechanism, the pressurized water spraying assembly comprises a pressurized water spraying pump and a spraying arm, the water inlet end of the pressurized water spraying pump is provided with a filter screen, the water outlet end of the pressurized water spraying pump is connected with the spraying arm with a hollow structure, the spraying arm is in a C shape, and a plurality of nozzles connected with the hollow structure are arranged on the spraying arm.

[0015] Preferably, the first arc-shaped arm and the second arc-shaped arm are each provided with a plurality of pressurized thrusters on the outer wall.

[0016] Preferably, the two ends of the threaded rod are installed in the mounting seat through bearing seats.

[0017] Preferably, the decoupling propulsion assembly comprises four decoupling thrusters and is symmetrically distributed in two groups, and each group of decoupling thrusters has an up-down difference.

[0018] Preferably, the control mechanism and the detection system are further provided with a fault diagnosis neural network model constructed by using a convolutional neural network and a multi-head self-attention mechanism, information collected by the sensing assembly is input into the fault diagnosis neural network model for analysis, and a cable detection result is output.

[0019] Based on the above technical solutions, the present application has the following advantages:

[0020] 1. The present application has superior operation stability and can improve the quality of detection data: through the C-shaped surrounding crawling structure, the robot and the cable form a stable contact support, thereby greatly reducing the direct disturbance of the parent robot shaking and ocean current on the detection platform, and providing a stable reference for optical, sonar and other sensors. This makes the collected images clear without jitter and artifacts, providing a high-quality data basis for subsequent intelligent diagnosis, greatly improving the fault recognition rate and positioning accuracy.

[0021] 2. The application has high efficiency and integrated pretreatment and detection capability: innovatively integrating the C-type reciprocating cleaning brush inside the annular operation chamber, realizing the "edge travel, edge cleaning, edge detection" pipeline operation mode. It can remove biological attachment and sediment on the surface of the cable before detection, providing a clean and unobstructed operation surface for visual and sonar detection, fundamentally solving the problem of failure detection caused by surface contamination, and significantly improving the effectiveness and efficiency of detection.

[0022] 3. The application has excellent complex environment passability and operation continuity: innovatively combining bionic inchworm obstacle crossing and decoupling propeller assisted relocation modes to build a comprehensive obstacle crossing strategy, wherein 1) for cable joints, repair sleeves and other inherent obstacles, the bionic inchworm "anchoring-stretching-anchoring-contracting" stepping mode is adopted to realize stable and low-consumption crossing without leaving the cable and collision; 2) for large obstacles or buried cable sections, the decoupling propeller is used to realize the overall crossing of the body after decoupling, and the cable tracking path planning is provided through visual and sonar images to realize recoupling crawling. This dual obstacle crossing capability improves the stability and emergency capability comprehensively, ensuring that the robot can realize long-distance and uninterrupted continuous detection in real and complex submarine cable environment, solving the core problem of the industry.

[0023] 4. The application has high intelligence and automation level: integrating multi-modal sensors and advanced fault diagnosis neural network model, it can realize real-time environmental sensing, fault identification and autonomous decision-making operation mode. Especially the active obstacle crossing based on visual servo and decoupling state path planning based on sonar search, which provides intelligent assistance for robot autonomous obstacle avoidance and navigation, greatly reduces the dependence on surface operators and operation difficulty.

[0024] 5. The application has optimized energy efficiency and reliability: the energy consumption of the surrounding crawling mode is much lower than that of the suspended operation mode against the water flow, significantly prolonging the operation time and range of a single deployment. The inchworm motion has low power consumption, and the bionic inchworm structure is only enabled in the inherent obstacle crossing mode. The tendon driving structure has high reliability and flexibility, ensuring the long-term stable operation of the system in high-pressure and corrosive underwater environment.

[0025] 6. The application has flexible functional expandability and application potential: it is easy to integrate additional modules such as eddy current detection, sediment cleaning and in-situ marking, which is not only suitable for submarine cables, but also widely applicable to the detection, maintenance and monitoring of other long linear underwater facilities such as underwater pipelines and cables, with broad market application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0026] The specific embodiments of the application will be further described in detail below with reference to the accompanying drawings.

[0027] Figure 1 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0028] Figure 2 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0029] Figure 3 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0030] Figure 4 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0031] Figure 5 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0032] Figure 6 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0033] Figure 7 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0034] Figure 8 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0035] Figure 9 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0036] Figure 10 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0037] Figure 11 Fig. 1 is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a bionic inchworm structure in an embodiment;

[0038] In the drawings, reference numerals are consistently used to designate components having similar functionalities across the several drawings.

[0039] 1, first surrounding mechanism; 10, rear end anchoring assembly; 101, anchoring foot unit; 1010, screw stepping motor; 1011, anchoring foot; 11, first driving assembly; 110, omni-directional driving wheel set; 1100, driving ball head; 1101, elastic support frame; 1102, wheel hub motor; 1103, driving wheel; 12, fixed arc-shaped arm; 13, movable arc-shaped arm; 14, hinge assembly; 140, pushing motor; 141, male arm; 142, female arm; 143, hinge shaft; 144, pushing screw; 145, hinge ball head; 146, hinge ball socket; 2, cleaning mechanism; 20, first arc-shaped arm; 21, second arc-shaped arm; 220, C-shaped cleaning brush; 221, connecting rod; 222, sleeve; 223, threaded rod; 224, cleaning motor; 225, mounting seat; 23, strip-shaped sliding hole; 240, water jet pump; 241, spraying arm; 2410, nozzle; 25, booster propeller; 3, telescopic bending mechanism; 30, flexible main body; 31, C-shaped framework; 311, guide hole; 32, telescopic motor; 33, tendon rope; 34, winch mechanism; 4, second surrounding mechanism; 40, front end anchoring assembly; 41, second driving assembly; 51, visual camera; 52, search sonar; 6, decoupling propeller; 7, control cabin. DETAILED DESCRIPTION

[0040] In order to more clearly illustrate the present application, the present application will be further described below in conjunction with preferred embodiments. It should be understood by those skilled in the art that the specific description below is illustrative rather than limiting, and should not limit the scope of protection of the present application.

[0041] In the description of the present application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0042] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be directly connected, or indirectly connected through an intermediate medium; it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0043] As Figures 1 to 9As shown, the embodiment provides a submarine cable fault detection robot based on a bionic inchworm structure, comprising: a first wrap-around mechanism 1, a cleaning mechanism 2, a telescopic bending mechanism 3 and a second wrap-around mechanism 4 connected in sequence, wherein,

[0044] The first wrap-around mechanism 1 and the second wrap-around mechanism 4 are both C-shaped, each comprising a fixed arc-shaped arm 12 and a movable arc-shaped arm 13 hinged by a hinge assembly 14, forming a semi-closed C-shaped structure with a movable opening. The hinge assembly 14 has a self-locking function and can be locked at any opening angle, ensuring that the movable arc-shaped arm 13 will not be accidentally opened due to external force during operation. The fixed arc-shaped arm 12 and the movable arc-shaped arm 13 are opened and closed based on the hinge assembly 14, and the opening and closing angle range is greater than 90 degrees, so that the robot can easily "embrace" or "release" cables of different diameters from the side. When the fixed arc-shaped arm 12 and the movable arc-shaped arm 13 are embraced, their inner diameter is slightly larger than the outer diameter of the cable, forming a ring-shaped operation chamber. The first wrap-around mechanism 1 and the second wrap-around mechanism 4 are respectively provided with a rear end anchoring assembly 10 and a front end anchoring assembly 40, the front end anchoring assembly 40 and the rear end anchoring assembly 10 each comprising a plurality of anchoring foot units 101; the first wrap-around mechanism 1 and the second wrap-around mechanism 4 are respectively provided with a first driving assembly 11 and a second driving assembly 41, the first driving assembly 11 and the second driving assembly 41 each comprising four omnidirectional drive wheel sets 110 arranged at the left front, right front, left rear and right rear positions of the inner wall of the frame, respectively, to form a symmetrical driving layout. Each set of omnidirectional drive wheel sets 110 is composed of a waterproof hub motor 1102 and a high-friction coefficient drive wheel 1103. The hub motor 1102 provides power and real-time feedback of the rotation speed through an encoder. Through precise control of the drive wheel 1103, the robot can advance and retreat along the cable axis direction on the surface of the cable. At the same time, by controlling the speed difference of the left and right wheel sets, the robot can realize micro-amplitude yaw rotation to compensate for the interference of ocean currents and sea waves. The outer wall of the first wrap-around mechanism 1 and the second wrap-around mechanism 4 is provided with a sensing assembly and a decoupling propulsion assembly, the sensing assembly comprising a vision camera 51 and a search sonar 52, and the decoupling propulsion assembly comprising a plurality of decoupling thrusters 6.

[0045] The cleaning mechanism 2 is C-shaped, as shown in Figure 7 、 8The cleaning mechanism 2 comprises a first arc-shaped arm 20 and a second arc-shaped arm 21 hinged through a hinge assembly 14, and a cleaning assembly arranged on each of the first arc-shaped arm 20 and the second arc-shaped arm 21, wherein a strip-shaped sliding hole 23 is formed on each of the first arc-shaped arm 20 and the second arc-shaped arm 21, and the cleaning assembly comprises a C-shaped cleaning brush 220, a connecting rod 221, a sleeve 222 and a cleaning motor 224, wherein an output shaft of the cleaning motor 224 is connected with a threaded rod 223 through a transmission gear box, one end of the connecting rod 221 is connected with the C-shaped cleaning brush 220, the other end of the connecting rod 221 is connected with the sleeve 222 through the strip-shaped sliding hole 23, the sleeve 222 is sleeved on the threaded rod 223, the sleeve 222 is threadedly connected with the threaded rod 223, and both ends of the threaded rod 223 are installed in a mounting seat 225 through bearing seats. The C-shaped cleaning brush is attached to the inner side of the first arc-shaped arm 20 or the second arc-shaped arm 21, the driving of the cleaning motor 224 drives the transmission gear box to rotate the threaded rod 223, the sleeve 222 threadedly connected with the threaded rod 223 moves along the cable axis in a reciprocating linear motion, and the C-shaped cleaning brush 220 connected with the sleeve 222 moves along the cable axis in a reciprocating linear motion, so that the surface of the cable is cleaned. After the robot grasps the cable and enters the embracing crawling, the cleaning mechanism 2 is started, and the robot continuously cleans the surface of the cable to ensure that the subsequent detection system obtains a clean detection surface.

[0046] The telescopic bending mechanism 3 comprises a flexible main body 30, a plurality of C-shaped skeletons 31 and N telescopic motors 32, the plurality of C-shaped skeletons 31 are embedded in the flexible main body 30, a plurality of groups of parallel distributed guide holes 311 are arranged on the plurality of C-shaped skeletons 31, the telescopic motor 32 is arranged on the cleaning mechanism 2, the telescopic motor 32 is connected with a winch mechanism 34, a tendon 33 wound on the winch mechanism 34 is fixedly connected with the second surrounding mechanism 4 after passing through each group of guide holes 311. The bionic inchworm structure comprises the rear end anchoring assembly 10 and the front end anchoring assembly 40 arranged on the first surrounding mechanism 1 and the second surrounding mechanism 4 respectively, and the telescopic bending mechanism 3, the rear end anchoring assembly 10 serves as a power source of the system and provides a rear fixed fulcrum for the obstacle crossing process, and the front end anchoring assembly 40 is used for determining the inchworm movement direction and providing a front fixed fulcrum. When an obstacle is encountered, the robot crosses the obstacle in a step-by-step manner of “locking the rear end anchoring assembly 10 → releasing the front end anchoring assembly 40 → stretching the telescopic bending mechanism 3 (achieved by the telescopic motor 32 in cooperation with the decoupling propeller 6 on the second surrounding mechanism 4) → locking the front end anchoring assembly 40 → releasing the rear end anchoring assembly 10 → shrinking the telescopic bending mechanism 3 (achieved by the telescopic motor 32)”. The bionic inchworm structure design enables the robot to stably cross small obstacles and has strong anti-disturbance ability. The flexible main body 30 is made of a high-elasticity corrosion-resistant composite material, and the C-shaped skeleton 31 is made of a lightweight high-strength alloy.

[0047] In one embodiment of the bionic inchworm structure-based subsea cable fault detection robot, referring to Figure 4 , 5 The hinge assembly 14 includes a male arm 141, a female arm 142, a hinge shaft 143, a pushing lead screw 144, and a pushing motor 140. The male arm 141 is hinged with the female arm 142 through the hinge shaft 143. The pushing motor 140 is arranged on the female arm 142. The output end of the pushing motor 140 is connected with one end of the pushing lead screw 144 through a shaft connector. The other end of the pushing lead screw 144 is connected with a hinge ball socket 146 on the male arm 141 through a hinge ball head 145. The hinge assembly 14 enables the fixed arc-shaped arm 12 and the movable arc-shaped arm 13 to open and close, with an opening and closing angle range greater than 90 degrees, so that the robot can easily “embrace” or “release” cables of different diameters from the side.

[0048] In one embodiment of the bionic inchworm structure-based subsea cable fault detection robot, the anchor foot units 101 are distributed at equal intervals, referring to Figure 4 The anchor foot unit 101 includes a lead screw stepping motor 1010 arranged on the outer wall of the fixed arc-shaped arm 12 or the movable arc-shaped arm 13. The output end of the lead screw stepping motor 1010 extends into the fixed arc-shaped arm 12 or the movable arc-shaped arm 13, and the end of the output end is provided with an anchor foot 1011. By controlling the forward and reverse rotation of the lead screw stepping motor 1010, the anchor foot 1011 is moved forward or backward along the axis direction, so as to realize the anchoring or loosening of the cable. An inner sealing sleeve and an outer sealing cover are arranged to seal and protect the lead screw stepping motor 1010. The outer sealing cover is embedded with a bearing device to provide additional radial support for the output end of the lead screw stepping motor 1010. The inner sealing sleeve is made of a corrosion-resistant, anti-aging and flexible composite material, which can seal while allowing the telescopic lead screw to freely extend and retract. The anchor foot 1011 is made of a material with moderate hardness and softness, has the characteristics of wear resistance and high friction coefficient, and is designed to provide sufficient locking friction while avoiding damage to the cable sheath.

[0049] In one embodiment of the bionic inchworm structure-based subsea cable fault detection robot, referring to Figure 9, several omni-directional drive wheel groups 110 are distributed equidistantly, the omni-directional drive wheel group 110 comprises a driving ball head 1100, an elastic support frame 1101, a wheel hub motor 1102, and a driving wheel 1103 integrated on the outer rotor of the wheel hub motor 1102, the wheel hub motor 1102 is installed on the elastic support frame 1101, and the top of the elastic support frame 1101 is connected with a driving ball socket on the fixed arc-shaped arm 12 or the movable arc-shaped arm 13 through the driving ball head 1100. Wherein, the wheel hub motor 1102 adopts an outer rotor type permanent magnet synchronous motor as power, the rotor shell of the wheel hub motor 1102 directly serves as a hub of the driving wheel 1103, the torque is directly transmitted to the driving wheel 1103, the volume is small, the structure is simplified, and the wheel hub motor 1102 has the characteristics of high efficiency and fast dynamic response. The rotor and the bearing are made of special stainless steel material, which ensures the reliability and durability of long-term operation in high-pressure and corrosive marine environment. The driving wheel 1103 adopts a composite material with high friction coefficient, seawater corrosion resistance and wear resistance, and the wheel surface of the driving wheel 1103 is designed as a concave arc surface matched with the cylindrical curved surface of the cable, so as to increase the contact area and provide stronger adhesion and traction. The wheel hub motor 1102 is installed on the elastic support frame 1101, which ensures that the driving wheel 1103 can effectively provide balanced and continuous driving force under complex cable surface conditions, and buffers the impact and vibration in operation. The elastic support frame 1101 is connected with the driving ball socket on the fixed arc-shaped arm 12 or the movable arc-shaped arm 13 through the driving ball head 1100, and has the ability of universal adjustment of the advancing direction. The encoder is built in the wheel hub motor 1102, which is used for feeding back the absolute position and rotating speed of the motor rotor, and provides real-time feedback for realizing the precise speed control and stroke positioning of the robot. By controlling all the wheel hub motors 1102 to rotate in the same direction and at the same speed, the robot can advance and retreat along the cable axis direction. By controlling the wheel hub motors 1102 on the left and right sides to generate a speed difference, the robot can make a small yaw rotation around the cable axis, so as to dynamically adjust the advancing direction, compensate for the lateral drift caused by external disturbances such as ocean current and sea wave, and ensure that the robot always stably advances along the cable axis direction. By accurately controlling the start-stop and rotating speed of each wheel hub motor 1102 and combining the feedback of the built-in encoder, the robot can accurately stop or change speed at a point on the cable, which is convenient for fixed-point and fine detection and cleaning of suspicious parts.

[0050] In the bionic inchworm structure-based seabed cable fault detection robot of one embodiment, referring to Figure 7 、 8The first arc-shaped arm 20 and the second arc-shaped arm 21 are provided with a pressurized water spraying assembly on the side wall close to the first surrounding mechanism 1, the pressurized water spraying assembly comprises a pressurized water spraying pump 240 and a spraying arm 241, the water inlet end of the pressurized water spraying pump 240 is provided with a filter screen to ensure the cleanliness of the washing water; the water outlet end of the pressurized water spraying pump 240 is connected with the spraying arm 241 having a hollow structure, the spraying arm 241 is in a C shape, and a plurality of nozzles 2410 connected with the hollow structure are arranged on the spraying arm 241 for spraying clean high-pressure water flow to the surface of the cable to assist in cleaning and washing away the foreign matters brushed away by the C-shaped cleaning brush 220. Further, referring to Figure 6 A plurality of pressurized thrusters 25 are arranged on the outer wall of the first arc-shaped arm 20 and the second arc-shaped arm 21 to assist in increasing the contact pressure between the cleaning brush surface and the cable surface and further adjusting the cleaning strength. Referring to Figure 6 、 7 、8, the C-shaped cleaning brush 220, the pressurized thruster 25 and the pressurized water spraying assembly jointly constitute a high-efficiency cooperative cleaning system, which removes the biological attachment and deposits on the surface of the cable in real time before detection to provide clean and unobstructed ideal working conditions for the imaging-based visual and sonar detection, thereby fundamentally solving the industry problem of missed detection and false detection caused by pollution and ensuring the accuracy and efficiency of the detection results.

[0051] In an embodiment of the subsea cable fault detection robot based on the bionic inchworm structure, referring to Figure 1 、 2 The decoupling propulsion assembly comprises four decoupling thrusters 6 and is symmetrically distributed in two groups, each group of decoupling thrusters 6 has an up-down position difference, and is arranged on the upper part of the robot body to provide thrust and torque for the robot in the decoupled state with the cable, wherein the four decoupling thrusters 6 of the upper layer are arranged in right front upper direction, left front upper direction, right rear upper direction and left rear upper direction respectively; the four decoupling thrusters 6 of the lower layer are all arranged in front-rear direction. If a large obstacle that cannot be crossed or a buried cable section is encountered, the front end anchoring assembly 40 and the rear end anchoring assembly 10 are completely loosened, the robot is completely separated from the cable, the feedback information of the visual camera 51 and the search sonar 52 is used for planning a path, the decoupling thrusters 6 push the robot to track the path to dive, and when the recoupling condition is reached, the robot is returned to continue the ring-shaped crawling detection until the detection is completed.

[0052] In an embodiment of the bionical inchworm structure-based seabed cable fault detection robot, a control mechanism and detection system are further included, wherein a fault diagnosis neural network model constructed using a convolutional neural network and a multi-head self-attention mechanism is arranged in the control mechanism and detection system, information collected by the perception component is input into the fault diagnosis neural network model for analysis, and a cable detection result is output. The perception component includes a visual camera 51 and a search sonar 52. The visual camera 51 is composed of two high-resolution industrial cameras to form a visual system, which can clearly capture the details of the seabed cable surface and provide high-quality image data for machine vision analysis. The two high-resolution industrial cameras have a pixel of 3840x2160 and are respectively arranged in their own alloy pressure-resistant sealed cabins. One of the two cameras is arranged at the top of the front end of the main body of the front end ring-shaped crawling structure and has a forward view, and the other is arranged at the top of the front and rear ends of the main body of the rear end ring-shaped crawling structure and has a backward view. The search sonar 52 includes two front-view search sonars, one of which is installed at the top of the front end of the main body of the front end ring-shaped crawling structure and has a detection beam pointing in the forward direction, and the other is installed at the top of the rear end of the main body of the rear end ring-shaped crawling structure and has a detection beam pointing in the backward direction. The working frequency of the sonar takes into account the resolution and penetration, which can effectively detect and identify the seabed terrain, obstacles and cable profile in front of the robot. For cables that are completely exposed or partially suspended, the sonar can provide clear morphological imaging. In addition, under the working condition that the cable is shallowly buried by seabed sediments, the sonar can still detect the buried position and general trend of the cable through low-frequency penetration technology based on the acoustic property difference between the buried object and the surrounding sediments. Comprehensive real-time spatial data are provided for the robot's autonomous obstacle avoidance, path planning in decoupled state and cable state evaluation.

[0053] The control mechanism and detection system are the central control of the underwater robot, responsible for coordinating and managing various subsystems to achieve efficient and stable operation of the robot. A high-performance multi-core processor and a real-time operating system are used to build the control mechanism and detection system. The control mechanism and detection system are packaged in the control cabin 7 using an alloy pressure-resistant sealed cabin. The control mechanism and detection system receive feedback data from various subsystems to make decisions and control. All control signals are connected to the robot main control cabin 7 through waterproof connection components to ensure stable communication in underwater environment.

[0054] The fault diagnosis neural network model adopts a hybrid structure of convolutional neural network and multi-head self-attention mechanism. The working state images of the submarine cable collected by the underwater robot in real time are extracted by the convolutional neural network to obtain multi-scale features, and the multi-head self-attention mechanism encoder is used to enhance the feature expression. The query mechanism and the decoding output the working state type of the submarine cable. A feature fusion module is adopted in one step, the sonar image features and the visual features are fused based on the timestamp alignment mechanism to enhance the discrimination of micro-faults, and then the state is distinguished by the decoder detection head. The fault detection system detects the working state of the cable at the current position in real time according to the current visual and sonar images. When a fault is detected, the detection head of the network outputs the defect type and confidence, and gives the global positioning coordinates measured by the current sensor in real time.

[0055] Further, the fault diagnosis neural network model, see Figure 10 、 11 , comprises a backbone network, a neck network and a detection head network, wherein the backbone network adopts a plurality of two-dimensional convolutional layers with a convolution kernel size of 3*3 and a step size of 2 to downsample the visual image (submarine cable image), thereby obtaining visual image feature maps of different scales. The sonar image features are preliminarily extracted by a two-dimensional convolutional layer with a convolution kernel size of 3*3, thereby obtaining sonar image feature maps. The visual image feature maps of different scales and the sonar image feature maps are further extracted by a feature extraction module and then output to the neck network. The feature extraction module adopts a joint frequency domain and spatial domain feature extraction module innovatively proposed, which is applied to the feature maps of different scales for joint feature extraction in the frequency domain and the spatial domain, thereby providing input features of different scales for the neck network.

[0056] The neck network mainly includes a multi-head self-attention module for scale-in feature interaction and a feature fusion module for cross-scale feature fusion. The multi-head self-attention module is mainly responsible for processing the top feature maps in the backbone network to enhance the expression ability of the features. The processed features are reshaped into vectors, then the features are interacted and transformed by the multi-head self-attention mechanism, and the processed features are reshaped back to two-dimensional space for subsequent cross-scale feature fusion. The feature fusion module aligns the features of different scales in size by upsampling and downsampling operations, and fuses the aligned feature maps by a reparameterization feature extraction method, thereby effectively integrating the feature information of different scales and improving the detection ability of different size targets.

[0057] The detection head network adopts a cross-attention decoder structure and combines a cross-union ratio perception query selection mechanism. The feature map transmitted by the neck network is flattened to obtain a two-dimensional matrix convenient for interaction with the query Q, and then the attention weight between the query Q and the two-dimensional matrix is calculated through the cross-attention mechanism. Further, the IoU perception query selection mechanism optimizes the selection of the best query through the intersection and union ratio information. Finally, the prediction head decodes and converts the optimized query into the final detection result, including the category and position information of the target.

[0058] The fault diagnosis neural network model detection process comprises:

[0059] S1: Obtain a submarine cable original data set, and preprocess the data set by using a data enhancement method to obtain a training set containing cable working state information.

[0060] S2: Train the neural network submarine cable fault detection model based on multi-modal information fusion provided by the application by using the training set obtained in step S1 to obtain a submarine cable fault diagnosis model.

[0061] S3: Adjust the hyperparameters of the training submarine cable fault detection network according to the model obtained by training in step S2 until the optimal fault diagnosis neural network model is obtained.

[0062] S4: Input the submarine cable visual image and sonar image collected by the underwater robot in real time into the optimal fault diagnosis neural network model obtained in step S3 to obtain the working state of each part of the collected cable, mark the part with a fault or potential threat, and send a warning to the early warning system.

[0063] The above only describes the preferred embodiment of the submarine cable fault detection robot based on the bionic inchworm structure disclosed in the application, and is not intended to limit the protection scope of the embodiments of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the embodiments of the present application shall be included in the protection scope of the embodiments of the present application.

[0064] It should be further noted that the terms “comprising”, “including”, or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or apparatus including a series of elements includes not only those elements, but also other elements not explicitly listed, or other elements inherent to such a process, method, article, or apparatus. Without more limitations, the element defined by the statement “including a” does not exclude the presence of additional identical elements in the process, method, article, or apparatus including the element.

Claims

1. A seabed cable fault detection robot based on a biomimetic inchworm structure, characterized by, The utility model relates to a cleaning robot, including: First loop mechanism, cleaning mechanism, telescopic bending mechanism and second loop mechanism are connected in turn, The first loop mechanism and second loop mechanism are all C-shaped, and all include fixed arc-shaped arm and movable arc-shaped arm articulated by hinge assembly, the first loop mechanism and second loop mechanism are provided with rear end anchoring assembly and front end anchoring assembly respectively, the front end anchoring assembly and rear end anchoring assembly all include several anchoring foot units, the first loop mechanism and second loop mechanism are provided with first drive assembly and second drive assembly respectively, the first drive assembly and second drive assembly all include several omni-directional drive wheel groups, the outer wall of the first loop mechanism and second loop mechanism is provided with sensing assembly and decoupling propulsion assembly, the sensing assembly includes visual camera and search sonar, and the decoupling propulsion assembly includes several decoupling thrusters, The cleaning mechanism is C-shaped, and includes first arc-shaped arm and second arc-shaped arm articulated by hinge assembly, the first arc-shaped arm and second arc-shaped arm are provided with cleaning assembly, the first arc-shaped arm and second arc-shaped arm are provided with strip sliding hole, the cleaning assembly includes C-shaped cleaning brush, connecting rod, sleeve and cleaning motor, the output shaft of cleaning motor is connected with threaded rod through transmission gear box, one end of connecting rod is connected with C-shaped cleaning brush, the other end of connecting rod passes through strip sliding hole and is connected with sleeve, the sleeve is sleeved on threaded rod, and the sleeve is threadedly connected with threaded rod, The telescopic bending mechanism includes flexible main body, several C-shaped skeletons and N telescopic motors, several C-shaped skeletons are embedded in flexible main body, several C-shaped skeletons are provided with N groups of parallel distributed guide holes, the telescopic motor is arranged on the cleaning mechanism, the telescopic motor is connected with winch mechanism, and the tendon rope wound on the winch mechanism is fixedly connected with the second loop mechanism after passing through each group of guide holes.

2. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, The hinge assembly includes male arm, female arm, hinge shaft, push screw and push motor, the male arm and female arm are articulated through hinge shaft, the push motor is arranged on the female arm, the output end of push motor is connected with one end of push screw through shaft connector, and the other end of push screw is connected with hinge ball socket on the male arm through hinge ball head.

3. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, Several anchoring foot units are distributed at equal distances, and the anchoring foot unit includes a lead screw stepper motor arranged on the outer wall of the fixed arc-shaped arm or movable arc-shaped arm, and the output end of the lead screw stepper motor extends into the fixed arc-shaped arm or movable arc-shaped arm and is provided with an anchoring foot at the end of the output end.

4. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, Several omni-directional drive wheel groups are distributed at equal distances, and the omni-directional drive wheel group includes drive ball head, elastic support frame, wheel hub motor, drive wheel, the drive wheel is integrated on the outer rotor of the wheel hub motor, the wheel hub motor is installed on the elastic support frame, and the top of the elastic support frame is connected with the drive ball socket on the fixed arc-shaped arm or movable arc-shaped arm through the drive ball head.

5. The biomimic inchworm structure based subsea cable fault detection robot of claim 4, wherein, The wheel surface of the drive wheel is a concave arc surface.

6. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, The first arc-shaped arm and the second arc-shaped arm are provided with a pressurized water spraying assembly on the side wall close to the first surrounding mechanism, the pressurized water spraying assembly comprises a pressurized water spraying pump and a spraying arm, the water inlet end of the pressurized water spraying pump is provided with a filter screen, the water outlet end of the pressurized water spraying pump is connected with the spraying arm with a hollow structure, the spraying arm is in a C shape, and a plurality of nozzles connected with the hollow structure are arranged on the spraying arm.

7. The biomimic inchworm structure based subsea cable fault detection robot of claim 1 or 6, wherein, The outer wall of the first arc-shaped arm and the second arc-shaped arm is provided with a plurality of pressurized thrusters.

8. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, The two ends of the threaded rod are installed in the mounting seat through bearing seats.

9. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, The decoupling propulsion assembly comprises four decoupling thrusters and is symmetrically distributed in two groups, and each group of decoupling thrusters has an up-down difference.

10. The biomimic inchworm structure based subsea cable fault detection robot of claim 1, wherein, Further comprising a control mechanism and a detection system, the control mechanism and the detection system are provided with a fault diagnosis neural network model constructed by using a convolutional neural network and a multi-head self-attention mechanism, information collected by the perception assembly is input into the fault diagnosis neural network model for analysis, and a cable detection result is output.

Images