Submarine cable fault detection robot based on bionic inchworm structure

Abstract

The invention discloses a submarine cable fault detection robot based on a bionic inchworm structure, and relates to the technical field of underwater robots. Comprising a first surrounding mechanism, a cleaning mechanism, a telescopic bending mechanism and a second surrounding mechanism which are connected in sequence, the cleaning mechanism integrated at the front end can automatically remove attachments on the surface of a cable before detection, and the detection accuracy is improved; the first surrounding mechanism, the telescopic bending mechanism and the second surrounding mechanism form a bionic inchworm structure, and the robot can step across inherent obstacles under the condition that the robot is not separated from a cable through the bionic inchworm structure; when crossing a large obstacle and a buried section of a cable, the robot decouples and advances through the decoupling propulsion assembly and tracks a cable path, and when a coupling condition is reached, the robot resets and continues to carry out embracing type crawling detection through the bionic inchworm structure, so that long-distance, uninterrupted and full-automatic fault detection on the complex submarine cable is realized. According to the invention, the quality of submarine cable detection data, the operation efficiency and the intelligent level are improved.

Description

Technical Field

[0001] This invention relates to the field of underwater robot technology, and in particular to a submarine cable fault detection robot based on a biomimetic inchworm structure. Background Technology

[0002] With the continuous development of transnational communications and offshore wind power, the safe operation and maintenance of underwater cable facilities faces enormous challenges. Traditional submarine cable inspection and maintenance mainly rely on manual diving operations, which have drawbacks such as low efficiency, high cost, and high risk. Especially in harsh environments such as deep-sea high pressure and low visibility, traditional methods are difficult to achieve reliable fault identification and rapid location.

[0003] Remotely operated underwater vehicles (ROVs) are increasingly being used in the laying, inspection, and maintenance of submarine cables. However, current ROV inspections of submarine cables primarily rely on imaging with a single, ordinary optical camera, followed by continuous, real-time visual observation. This method of ROVs hovering near the cable and observing via the onboard optical camera has significant drawbacks. First, underwater currents and waves can cause the inspection platform to sway, resulting in targets being outside the camera's field of view, blurred images, and distorted data. Second, hovering operations are energy-intensive, have short endurance, and are difficult to control. Furthermore, biological attachments and sediment covering the cable surface can severely obstruct vision, leading to missed or false faults. On the other hand, traditional attached crawling inspection equipment cannot continue its progress when encountering obstacles such as cable joints, repair sleeves, or when parts of the cable are buried, forcing the inspection operation to be interrupted.

[0004] Therefore, there is an urgent need for an intelligent submarine cable fault detection robot that can operate stably, efficiently, and continuously, and overcome the aforementioned unfavorable conditions. Summary of the Invention

[0005] To address the aforementioned technical issues, this invention proposes a submarine cable fault detection robot based on a biomimetic inchworm structure. Through its C-shaped crawling structure, autonomous cleaning capabilities, and multi-mode obstacle-crossing ability, it can achieve long-distance, uninterrupted, fully automated continuous detection and fault diagnosis operations on the seabed.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] A submarine cable fault detection robot based on a biomimetic inchworm structure includes: a first circling mechanism, a cleaning mechanism, a telescopic bending mechanism, and a second circling mechanism connected in sequence. Both the first and second circling mechanisms are C-shaped and include a fixed arc-shaped arm and a movable arc-shaped arm hinged by a hinge assembly. The first and second circling mechanisms are respectively equipped with a rear-end anchoring assembly and a front-end anchoring assembly, each including several anchoring foot units. The first and second circling mechanisms are respectively equipped with a first driving assembly and a second driving assembly, each including several omnidirectional drive wheel sets. The outer walls of both the first and second circling mechanisms are equipped with a sensing assembly and a decoupling propulsion assembly. The sensing assembly includes a visual camera and a search sonar, and the decoupling propulsion assembly includes several decoupling thrusters.

[0008] The cleaning mechanism is C-shaped and includes a first arc-shaped arm and a second arc-shaped arm hinged together by a hinge assembly. Each of the first and second arc-shaped arms is equipped with a cleaning component. Each of the first and second arc-shaped arms has a strip-shaped sliding hole. The cleaning component includes a C-shaped cleaning brush, a connecting rod, a sleeve, and a cleaning motor. The output shaft of the cleaning motor is connected to a threaded rod through a transmission gearbox. One end of the connecting rod is connected to the C-shaped cleaning brush, and the other end of the connecting rod passes through the strip-shaped sliding hole and is connected to the sleeve. The sleeve is fitted onto the threaded rod, and the sleeve is threadedly connected to the threaded rod.

[0009] The telescopic bending mechanism includes a flexible body, several C-shaped frames, and N telescopic motors. Several C-shaped frames are embedded in the flexible body, and N sets of parallel guide holes are provided on the several C-shaped frames. The telescopic motors are provided on the cleaning mechanism and are connected to the winch mechanism. The tendon rope wound on the winch mechanism passes through each set of guide holes and is fixedly connected to the second circling mechanism.

[0010] Preferably, the hinge assembly includes a male arm, a female arm, a hinge shaft, a push screw, and a push motor. The male arm and the female arm are hinged together by the hinge shaft. The push motor is mounted on the female arm. The output end of the push motor is connected to one end of the push screw through a shaft connector. The other end of the push screw is connected to the hinge ball joint on the male arm through a hinge ball joint.

[0011] Preferably, a plurality of anchoring foot units are distributed at equal distances. Each anchoring foot unit includes a lead screw stepper motor disposed on the outer wall of the fixed arc arm or the movable arc arm. The output end of the lead screw stepper motor extends into the fixed arc arm or the movable arc arm, and the end of the output end is provided with an anchoring foot.

[0012] Preferably, several omnidirectional drive wheel sets are distributed at equal distances; the omnidirectional drive wheel set includes a drive ball joint, an elastic support frame, a hub motor, and a drive wheel. The drive wheel is integrated on the outer rotor of the hub motor, the hub motor is mounted on the elastic support frame, and the top of the elastic support frame is connected to a drive ball socket on a fixed arc arm or a movable arc arm through the drive ball joint.

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

[0014] Preferably, both the first and second arc-shaped arms are provided with pressurized water spraying components on their sidewalls near the direction of the first surrounding mechanism. The pressurized water spraying components include a pressurized water spraying pump and a spraying arm. The inlet end of the pressurized water spraying pump is provided with a filter screen, and the outlet end of the pressurized water spraying pump is connected to a spraying arm with a hollow structure. The spraying arm is C-shaped and is provided with a plurality of nozzles connected to the hollow structure.

[0015] Preferably, a plurality of booster propellers are provided on the outer walls of both the first and second arc-shaped arms.

[0016] Preferably, both ends of the threaded rod are mounted in the mounting base via bearing seats.

[0017] Preferably, the decoupled propulsion assembly includes four decoupled thrusters symmetrically distributed in pairs, with an upper and lower position difference between each pair of decoupled thrusters.

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

[0019] Based on the above technical solution, the beneficial effects of the present invention are:

[0020] 1. This invention offers superior operational stability and improves the quality of detection data: The C-shaped, spiraling crawling structure creates stable contact support between the robot and the cable, significantly reducing the direct disturbance of the detection platform caused by the swaying of the parent robot and ocean currents, providing a stable reference for optical, sonar, and other sensors. This results in clear, jitter-free, and artifact-free images, providing a high-quality data foundation for subsequent intelligent diagnostics and greatly improving fault identification rate and positioning accuracy.

[0021] 2. This invention possesses highly efficient integrated pre-treatment and inspection capabilities: It innovatively integrates a C-type reciprocating cleaning brush into the annular working chamber, realizing a streamlined operation mode of "moving, cleaning, and inspecting simultaneously." It can instantly remove biological attachments and deposits from the cable surface before inspection, providing a clean and unobstructed working surface for visual and sonar inspections. This fundamentally solves the problem of missed or false detections of faults caused by surface contamination, significantly improving the effectiveness and efficiency of inspections.

[0022] 3. This invention boasts superior versatility and operational continuity in complex environments: it uniquely integrates two modes—biomimetic inchworm obstacle crossing and decoupled thruster-assisted repositioning—to construct a comprehensive obstacle-crossing strategy. Specifically: 1) For inherent obstacles such as cable joints and repair sleeves, it employs a biomimetic inchworm-like "anchor-extend-anchor-retract" stepping pattern, achieving stable and low-energy crossing without detaching from the cable or experiencing collisions; 2) For large obstacles or buried cable sections, it utilizes multi-directionally arranged decoupled thrusters to achieve overall crossing after decoupling, and provides cable tracking path planning through visual and sonar images to achieve recoupling and crawling. This dual obstacle-crossing capability comprehensively improves stability and emergency response capabilities, ensuring the robot can perform long-distance, uninterrupted continuous inspection in real, complex submarine cable environments, solving a core industry shortcoming.

[0023] 4. This invention boasts a high level of intelligence and automation: integrating multimodal sensors and an advanced fault diagnosis neural network model, it can perceive the environment in real time, identify faults, and autonomously decide on operating modes. In particular, the combination of visual servoing-based active obstacle crossing and sonar-based decoupled state path planning, based on the fault diagnosis neural network model, provides intelligent assistance for the robot's autonomous obstacle avoidance and navigation, significantly reducing reliance on surface operators and the operational difficulty.

[0024] 5. This invention features optimized energy efficiency and reliability: the energy consumption of the circling crawling mode is significantly lower than that of the continuous water-flow-resistant suspension operation mode, significantly extending the operation time and range of a single deployment. The inchworm-like movement consumes low power, and the biomimetic inchworm structure is only activated in the mode of crossing inherent obstacles. The tendon-driven structure has high reliability and flexibility, ensuring long-term stable operation of the system in high-pressure, corrosive underwater environments.

[0025] 6. This invention has flexible functional expandability and application potential: it is easy to integrate additional modules such as eddy current testing, sediment cleaning, and in-situ marking. It is not only suitable for submarine cables, but can also be widely used in the inspection, maintenance and monitoring of other long linear underwater facilities such as underwater pipelines and cables, and has broad market application prospects. Attached Figure Description

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

[0027] Figure 1 This is a schematic diagram of the overall structure of a submarine cable fault detection robot based on a biomimetic inchworm structure in one embodiment;

[0028] Figure 2 This is a frontal view of a submarine cable fault detection robot based on a biomimetic inchworm structure in one embodiment;

[0029] Figure 3 This is a bottom view of a submarine cable fault detection robot based on a biomimetic inchworm structure in one embodiment;

[0030] Figure 4 This is a partial structural schematic diagram of the second surrounding mechanism in one embodiment;

[0031] Figure 5 This is a structural schematic diagram of a hinge assembly in one embodiment;

[0032] Figure 6 This is a schematic diagram of the disassembled structure of a submarine cable fault detection robot based on a biomimetic inchworm structure in one embodiment;

[0033] Figure 7 This is a schematic diagram of the C-shaped cleaning brush in one embodiment;

[0034] Figure 8 This is a schematic diagram of the cleaning mechanism in one embodiment;

[0035] Figure 9 This is a schematic diagram of the omnidirectional drive wheel assembly in one embodiment;

[0036] Figure 10 This is a schematic diagram of the structure of a fault diagnosis neural network model in one embodiment;

[0037] Figure 11 This is a schematic diagram of the feature extraction module in one embodiment;

[0038] In the figure, the attached figures are labeled as follows:

[0039] 1. First surrounding mechanism; 10. Rear anchoring assembly; 101. Anchoring foot unit; 1010. Lead screw stepper motor; 1011. Anchoring foot; 11. First drive assembly; 110. Omnidirectional drive wheel set; 1100. Drive ball joint; 1101. Elastic support frame; 1102. Hub motor; 1103. Drive wheel; 12. Fixed arc arm; 13. Movable arc arm; 14. Hinge assembly; 140. Push motor; 141. Male arm; 142. Female arm; 143. Hinge shaft; 144. Push lead screw; 145. Hinge ball joint; 146. Hinge ball socket; 2. Cleaning mechanism; 20. First arc arm; 21. Second arc arm; 220. C-shaped cleaning brush; 221. Connecting rod; 222. Sleeve; 223. Threaded rod; 224. Cleaning motor; 225. Mounting base; 23. Strip-shaped sliding hole; 240. Booster water pump; 241. Spray arm; 2410. Nozzle; 25. Booster thruster; 3. Telescopic bending mechanism; 30. Flexible body; 31. C-shaped frame; 311. Guide hole; 32. Telescopic motor; 33. Tendon rope; 34. Winch mechanism; 4. Second surround mechanism; 40. Front anchoring assembly; 41. Second drive assembly; 51. Vision camera; 52. Search sonar; 6. Decoupling thruster; 7. Control cabin. Detailed Implementation

[0040] To more clearly illustrate the present invention, the following description, in conjunction with preferred embodiments, further clarifies the invention. Those skilled in the art should understand that the specific descriptions below are illustrative rather than restrictive, and should not be construed as limiting the scope of protection of the present invention.

[0041] In the description of this invention, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0042] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0043] like Figures 1 to 9As shown, this embodiment provides a submarine cable fault detection robot based on a biomimetic inchworm structure, comprising: a first encircling mechanism 1, a cleaning mechanism 2, a telescopic bending mechanism 3, and a second encircling mechanism 4 connected in sequence, wherein...

[0044] Both the first surrounding mechanism 1 and the second surrounding mechanism 4 are C-shaped, each including a fixed arc-shaped arm 12 and a movable arc-shaped arm 13 hinged together by a hinge assembly 14, forming a semi-enclosed C-shaped structure with a movable opening. The hinge assembly 14 has a self-locking function, capable of locking at any opening angle to ensure that the movable arc-shaped arm 13 will not be accidentally opened by external force during operation. The fixed arc-shaped arm 12 and the movable arc-shaped arm 13 open and close based on the hinge assembly 14, with an opening angle range greater than 90 degrees, so that the robot can easily "hug" or "release" cables of different diameters from the side. When the fixed arc-shaped arm 12 and the movable arc-shaped arm 13 are engaged, their inner diameter is slightly larger than the outer diameter of the cable, forming an annular working chamber. The first surrounding mechanism 1 and the second surrounding mechanism 4 are respectively equipped with a rear anchoring component 10 and a front anchoring component 40. Both the front anchoring component 40 and the rear anchoring component 10 include several anchoring foot units 101. The first surrounding mechanism 1 and the second surrounding mechanism 4 are respectively equipped with a first drive component 11 and a second drive component 41. Both the first drive component 11 and the second drive component 41 include four omnidirectional drive wheel sets 110, respectively arranged at the left front, right front, left rear, and right rear positions on the inner wall of the frame, forming a symmetrical drive layout. Each omnidirectional drive wheel set 110 consists of a waterproof hub motor 1102 and a high-friction drive wheel 1103. The hub motor 1102 provides power and provides real-time speed feedback through an encoder. Through precise control of the drive wheels 1103, the robot can move forward and backward along the cable axis on the cable surface. Simultaneously, by controlling the speed difference between the left and right wheel sets, the robot can achieve slight yaw rotation, compensating for interference from ocean currents and waves. The outer walls of the first surrounding mechanism 1 and the second surrounding mechanism 4 are each provided with a sensing component and a decoupling propulsion component. The sensing component includes a visual camera 51 and a search sonar 52, and the decoupling propulsion component includes a plurality of decoupling propellers 6.

[0045] The cleaning mechanism 2 is C-shaped, see [reference] Figure 7 , 8The device includes a first arc-shaped arm 20 and a second arc-shaped arm 21 hinged together by a hinge assembly 14. Both the first arc-shaped arm 20 and the second arc-shaped arm 21 are equipped with cleaning components. Both the first arc-shaped arm 20 and the second arc-shaped arm 21 have strip-shaped sliding holes 23. The cleaning components include a C-shaped cleaning brush 220, a connecting rod 221, a sleeve 222, and a cleaning motor 224. The output shaft of the cleaning motor 224 is connected to a threaded rod 223 through a transmission gearbox. One end of the connecting rod 221 is connected to the C-shaped cleaning brush 220, and the other end of the connecting rod 221 passes through the strip-shaped sliding hole 23 and is connected to the sleeve 222. The sleeve 222 is sleeved on the threaded rod 223 and is threadedly connected to the threaded rod 223. Both ends of the threaded rod 223 are installed in the mounting base 225 through bearing seats. The C-shaped cleaning brush is fitted onto the inner side of the first arc-shaped arm 20 or the second arc-shaped arm 21. Driven by the cleaning motor 224, the transmission gearbox rotates the threaded rod 223, causing the sleeve 222, threadedly connected to the threaded rod 223, to reciprocate linearly along the cable axis. This, in turn, causes the C-shaped cleaning brush 220, connected to the sleeve 222, to reciprocate linearly along the cable axis, thus cleaning the cable surface. After the robot grasps the cable and enters a wraparound crawling motion, the cleaning mechanism 2 is activated. The robot continuously cleans the cable surface to ensure that the subsequent inspection system obtains a clean inspection surface.

[0046] The telescopic bending mechanism 3 includes a flexible body 30, several C-shaped skeletons 31, and N telescopic motors 32. The C-shaped skeletons 31 are embedded within the flexible body 30, and N sets of parallel guide holes 311 are provided on each C-shaped skeleton 31. The telescopic motors 32 are mounted on the cleaning mechanism 2 and connected to a winch mechanism 34. The tendon rope 33 wound on the winch mechanism 34 passes through each set of guide holes 311 and is fixedly connected to the second circling mechanism 4. The biomimetic inchworm structure includes a rear anchoring component 10 and a front anchoring component 40 respectively mounted on the first circling mechanism 1 and the second circling mechanism 4, as well as the telescopic bending mechanism 3. The rear anchoring component 10 serves as the system's power source and provides a rear fixed support point during obstacle crossing, while the front anchoring component 40 determines the inchworm's movement direction and provides a front fixed support point. When encountering obstacles, the robot traverses them using a stepping mechanism: "locking the rear anchoring component 10 → releasing the front anchoring component 40 → extending the telescopic bending mechanism 3 (achieved by the telescopic motor 32 in conjunction with the decoupled thruster 6 on the second encircling mechanism 4) → locking the front anchoring component 40 → releasing the rear anchoring component 10 → retracting the telescopic bending mechanism 3 (achieved by the telescopic motor 32)." This biomimetic inchworm-like structural design enables the robot to stably overcome small obstacles and provides strong resistance to disturbances. The flexible body 30 is made of highly elastic and corrosion-resistant composite material, while the C-shaped frame 31 is made of lightweight, high-strength alloy.

[0047] In one embodiment of a submarine cable fault detection robot based on a biomimetic inchworm structure, see [link to relevant documentation]. Figure 4 , 5 The hinge assembly 14 includes a male arm 141, a female arm 142, a hinge shaft 143, a push screw 144, and a push motor 140. The male arm 141 and the female arm 142 are hinged together by the hinge shaft 143. The push motor 140 is mounted on the female arm 142. The output end of the push motor 140 is connected to one end of the push screw 144 via a shaft connector. The other end of the push screw 144 is connected to the hinge ball joint 146 on the male arm 141 via a hinge ball joint 145. The hinge assembly 14 enables the fixed arc arm 12 and the movable arc arm 13 to open and close, with an opening and closing angle range greater than 90 degrees, allowing the robot to easily "hug" or "release" cables of different diameters from the side.

[0048] In one embodiment of a submarine cable fault detection robot based on a biomimetic inchworm structure, several anchoring foot units 101 are evenly distributed, see [reference needed]. Figure 4 The anchoring foot unit 101 includes a lead screw stepper motor 1010 mounted on the outer wall of the fixed arc arm 12 or the movable arc arm 13. The output end of the lead screw stepper motor 1010 extends into the fixed arc arm 12 or the movable arc arm 13, and an anchoring foot 1011 is provided at the end of the output end. By controlling the forward and reverse operation of the lead screw stepper motor 1010, the anchoring foot 1011 moves forward or backward along the axial direction, thereby anchoring or releasing the cable. An inner sealing sleeve and an outer sealing cover are provided to seal and protect the lead screw stepper motor 1010. The outer sealing cover has an embedded bearing device to provide additional radial support for the output end of the lead screw stepper motor 1010; the inner sealing sleeve is made of a corrosion-resistant, anti-aging, and tough composite material, allowing the telescopic lead screw to extend and retract freely while sealing. The anchoring foot 1011 is made of a material with moderate hardness and softness, possessing wear-resistant and high friction coefficient characteristics, designed to provide sufficient locking friction while avoiding damage to the cable sheath.

[0049] In one embodiment of a submarine cable fault detection robot based on a biomimetic inchworm structure, see [link to relevant documentation]. Figure 9Several omnidirectional drive wheel sets 110 are evenly distributed. Each omnidirectional drive wheel set 110 includes a drive ball joint 1100, an elastic support frame 1101, a hub motor 1102, and a drive wheel 1103. The drive wheel 1103 is integrated on the outer rotor of the hub motor 1102. The hub motor 1102 is mounted on the elastic support frame 1101. The top of the elastic support frame 1101 is connected to a drive ball socket on a fixed arc-shaped arm 12 or a movable arc-shaped arm 13 via the drive ball joint 1100. The hub motor 1102 uses an external rotor type permanent magnet synchronous motor as its power source. The rotor housing of the hub motor 1102 directly serves as the hub of the drive wheel 1103, and torque is directly transmitted to the drive wheel 1103. It features small size, simplified structure, high efficiency, and rapid dynamic response. The rotor and bearings are made of special stainless steel to ensure reliability and durability during long-term operation in high-pressure, corrosive marine environments. The drive wheel 1103 is made of a high-friction, seawater-corrosion-resistant, and wear-resistant composite material. The wheel surface of the drive wheel 1103 is designed as a concave arc surface that matches the cylindrical curved surface of the cable to increase the contact area and provide stronger adhesion and traction. The hub motor 1102 is mounted on the elastic support frame 1101 to ensure that the drive wheel 1103 can effectively provide balanced and continuous driving force under complex cable surface conditions, while also buffering impacts and vibrations during operation. The upper end of the elastic support frame 1101 is connected to the drive ball joint on the fixed arc arm 12 or the movable arc arm 13 via the drive ball head 1100, providing omnidirectional adjustment of the travel direction. The encoder is built into the hub motor 1102 to provide feedback on the absolute position and speed of the motor rotor, providing real-time feedback for precise speed control and stroke positioning of the robot. By controlling all hub motors 1102 to rotate in the same direction and at the same speed, the robot can move forward and backward along the cable axis. By controlling the speed difference between the left and right hub motors 1102, the robot can achieve slight yaw rotation around the cable axis, dynamically adjusting its direction of travel and compensating for lateral drift caused by external disturbances such as ocean currents and waves, ensuring the robot always travels stably along the cable axis. Through precise control of the start, stop, and speed of each hub motor 1102, combined with feedback from the built-in encoder, the robot can stop or change speed at precise points on the cable, facilitating targeted and detailed inspection and cleaning of suspicious areas.

[0050] In one embodiment of a submarine cable fault detection robot based on a biomimetic inchworm structure, see [link to relevant documentation]. Figure 7 , 8Both the first arc-shaped arm 20 and the second arc-shaped arm 21 are equipped with pressurized water spray assemblies on their sidewalls near the first surrounding mechanism 1. Each pressurized water spray assembly includes a pressurized water spray pump 240 and a spray arm 241. The inlet end of the pressurized water spray pump 240 is equipped with a filter screen to ensure the cleanliness of the rinsing water. The outlet end of the pressurized water spray pump 240 is connected to a spray arm 241 with a hollow structure. The spray arm 241 is C-shaped and has several nozzles 2410 connected to the hollow structure for spraying clean, high-pressure water onto the cable surface to assist in cleaning and flushing away foreign matter removed by the C-shaped cleaning brush 220. Further, see... Figure 6 Several booster propellers 25 are provided on the outer walls of both 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, thereby further adjusting the cleaning intensity. See also Figure 6 , 7 The system, consisting of a C-shaped cleaning brush 220, a booster propeller 25, and a booster water spray assembly, forms a highly efficient collaborative cleaning system. This system removes biological deposits, sediments, and other dirt from the cable surface before inspection, providing clean and unobstructed ideal operating conditions for imaging-based visual and sonar inspections. This fundamentally solves the industry problem of missed or false detections caused by contamination, ensuring the accuracy and high efficiency of the inspection results.

[0051] In one embodiment of a submarine cable fault detection robot based on a biomimetic inchworm structure, see [link to relevant documentation]. Figure 1 , 2 The decoupling propulsion assembly includes four decoupling thrusters 6, symmetrically distributed in pairs. Each pair of decoupling thrusters 6 has a vertical difference between them and is arranged on the upper part of the robot body. They provide thrust and torque to the robot in a decoupled state from the cable. The four upper-layer decoupling thrusters 6 are arranged in the right-front-upward, left-front-upward, right-rear-upward, and left-rear-upward directions, respectively; the four lower-layer decoupling thrusters 6 are all set in the forward-backward direction. If a large, insurmountable obstacle or a buried section of the cable is encountered, the front anchoring assembly 40 and the rear anchoring assembly 10 are fully released, and the robot completely detaches from the cable. Using feedback information from the vision camera 51 and the search sonar 52, the robot plans a path, and the decoupling thrusters 6 propel the robot to follow the path and crawl. When the recoupling conditions are met, the robot returns to its original position and continues its circling crawling detection until the detection ends.

[0052] In one embodiment of the submarine cable fault detection robot based on a biomimetic inchworm structure, a control mechanism and a detection system are also included. The control mechanism and detection system are equipped with a fault diagnosis neural network model constructed using a convolutional neural network and a multi-head self-attention mechanism. Information collected by the sensing components is input into the fault diagnosis neural network model for analysis, and the cable detection results are output. The sensing components include a vision camera 51 and a search sonar 52. The vision camera 51 consists of a vision system composed of two high-resolution industrial cameras, capable of clearly capturing details on the surface of the submarine cable and providing high-quality image data for machine vision analysis. Both high-resolution industrial cameras have a resolution of 3840×2160 pixels and are housed in their respective alloy pressure-resistant sealed chambers. One camera is positioned at the top front end of the front-end annular crawling structure, facing forward, while the other is positioned at the top front and rear ends of the rear annular crawling structure, facing backward. The search sonar 52 includes two forward-looking search sonars. One is installed at the top front end of the front annular crawling structure, with its detection beam pointing in the forward direction; the other is installed at the top rear end of the rear annular crawling structure, with its detection beam pointing in the backward direction. This sonar operates at a frequency that balances resolution and penetration, effectively detecting and identifying seabed topography, obstacles, and cable outlines in front of the robot. For fully exposed or partially suspended cables, the sonar provides clear morphological imaging. Furthermore, even when cables are shallowly buried by seabed sediments, the sonar can still detect the cable's location and approximate direction using low-frequency penetration technology, based on the acoustic differences between the burial site and the surrounding sediment. This provides comprehensive real-time spatial data for the robot's autonomous obstacle avoidance, path planning in decoupled states, and cable condition assessment.

[0053] The control mechanism and detection system is the central control hub of the underwater robot, responsible for coordinating and managing various subsystems to ensure efficient and stable operation. It is built using a high-performance multi-core processor and a real-time operating system. The control mechanism and detection system are encapsulated in a pressure-resistant alloy sealed chamber within the control compartment 7. The control mechanism and detection system makes decisions and performs control by receiving feedback data from each subsystem. All control signals are connected to the robot's main control compartment 7 via waterproof connection components, ensuring stable communication in the underwater environment.

[0054] The fault diagnosis neural network model employs a hybrid structure of convolutional neural networks and multi-head self-attention mechanisms. Real-time images of the submarine cable's operational status acquired by an underwater robot are processed by the convolutional neural network to extract multi-scale features, which are then enhanced by a multi-head self-attention encoder. The resulting output, after querying and decoding, indicates the submarine cable's operational status type. A feature fusion module is then used, incorporating sonar image features with visual features based on a timestamp alignment mechanism to enhance the discriminative power of minor faults. Finally, a decoder head is used for status determination. The fault detection system continuously monitors the cable's operational status at its current location based on current visual and sonar images. When a fault is detected, the network's detection head outputs the defect type and confidence level, along with the global positioning coordinates measured by the current sensors.

[0055] For further information on fault diagnosis neural network models, please refer to [link / reference]. Figure 10 , 11 The system comprises a backbone network, a neck network, and a detection head network. The backbone network downsamples the visual image (submarine cable image) using multiple 2D convolutional layers with 3×3 kernels and a stride of 2, obtaining visual image feature maps at different scales. Preliminary feature extraction is performed on the sonar image features using 2D convolutional layers with 3×3 kernels, yielding sonar image feature maps. The visual and sonar image feature maps at different scales are further processed by a feature extraction module before being output to the neck network. This feature extraction module employs an innovative joint frequency-spatial domain feature extraction module, applying it to feature maps at different scales to perform joint frequency-spatial domain feature extraction, providing the neck network with input features at different scales.

[0056] The neck network mainly consists of two parts: a multi-head self-attention module for intra-scale feature interaction and a cross-scale feature fusion module. The multi-head self-attention module is primarily responsible for processing the top feature maps in the backbone network to enhance feature expressiveness. It flattens the two-dimensional feature maps into vectors, then performs feature interaction and transformation through a multi-head self-attention mechanism. The processed features are then reshaped back into two-dimensional space for subsequent cross-scale feature fusion. The feature fusion module aligns features at different scales through upsampling and downsampling operations, and fuses the aligned feature maps using a reparameterized feature extraction method. This effectively integrates feature information from different scales, improving the detection capability for targets of different sizes.

[0057] The detection head network employs a cross-attention decoder structure combined with an IoU-aware query selection mechanism. The feature map input to the neck network is flattened into a two-dimensional matrix that facilitates interaction with the query Q. Then, the attention weights between the query Q and the neck network are calculated using the cross-attention mechanism. Furthermore, the IoU-aware query selection mechanism optimizes the selection of the best query using IoU information. Finally, the prediction head decodes the optimized query and converts it into the final detection result, including the target's category and location information.

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

[0059] S1: Obtain the original dataset of submarine cables and preprocess the dataset using data augmentation methods to obtain a training set containing cable operating status information.

[0060] S2: Using the training set obtained in step S1, the neural network submarine cable fault detection model based on multimodal information fusion provided by this invention is trained to obtain a submarine cable fault diagnosis model.

[0061] S3: Based on the model trained in step S2, adjust the hyperparameters of the submarine cable fault detection network until the optimal fault diagnosis neural network model is obtained.

[0062] S4: Input the visual and sonar images of the submarine cable collected in real time by the underwater robot into the optimal fault diagnosis neural network model obtained in step S3 to obtain the working status of each part of the cable. Mark the parts with faults or potential threats and issue a warning to the early warning system.

[0063] The above description is merely a preferred embodiment of the submarine cable fault detection robot based on a biomimetic inchworm structure disclosed in this invention, and is not intended to limit the scope of protection of the embodiments in this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments in this specification should be included within the scope of protection of the embodiments in this specification.

[0064] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A submarine cable fault detection robot based on a biomimetic inchworm structure, characterized in that, include: The first surrounding mechanism, the cleaning mechanism, the telescopic bending mechanism, and the second surrounding mechanism are connected in sequence, wherein... Both the first and second encircling mechanisms are C-shaped and include a fixed arc-shaped arm and a movable arc-shaped arm hinged by a hinge assembly. The first and second encircling mechanisms are respectively provided with a rear-end anchoring assembly and a front-end anchoring assembly, each including several anchoring foot units. The first and second encircling mechanisms are respectively provided with a first driving assembly and a second driving assembly, each including several omnidirectional drive wheel sets. The outer walls of both the first and second encircling mechanisms are provided with a sensing assembly and a decoupling propulsion assembly. The sensing assembly includes a visual camera and a search sonar, and the decoupling propulsion assembly includes several decoupling propellers. The cleaning mechanism is C-shaped and includes a first arc-shaped arm and a second arc-shaped arm hinged together by a hinge assembly. Each of the first and second arc-shaped arms is equipped with a cleaning component. Each of the first and second arc-shaped arms has a strip-shaped sliding hole. The cleaning component includes a C-shaped cleaning brush, a connecting rod, a sleeve, and a cleaning motor. The output shaft of the cleaning motor is connected to a threaded rod through a transmission gearbox. One end of the connecting rod is connected to the C-shaped cleaning brush, and the other end of the connecting rod passes through the strip-shaped sliding hole and is connected to the sleeve. The sleeve is fitted onto the threaded rod, and the sleeve is threadedly connected to the threaded rod. The telescopic bending mechanism includes a flexible body, several C-shaped frames, and N telescopic motors. Several C-shaped frames are embedded in the flexible body, and N sets of parallel guide holes are provided on the several C-shaped frames. The telescopic motors are provided on the cleaning mechanism and are connected to the winch mechanism. The tendon rope wound on the winch mechanism passes through each set of guide holes and is fixedly connected to the second circling mechanism.

2. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, The hinge assembly includes a male arm, a female arm, a hinge shaft, a lead screw, and a drive motor. The male arm and the female arm are hinged together by the hinge shaft. The drive motor is mounted on the female arm. The output end of the drive motor is connected to one end of the lead screw via a shaft connector. The other end of the lead screw is connected to the hinge ball joint on the male arm via a hinge ball joint.

3. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, Several anchoring foot units are distributed at equal intervals. Each anchoring foot unit includes a lead screw stepper motor disposed on the outer wall of the fixed arc arm or the movable arc arm. The output end of the lead screw stepper motor extends into the fixed arc arm or the movable arc arm, and the end of the output end is provided with an anchoring foot.

4. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, Several omnidirectional drive wheel sets are evenly distributed; each omnidirectional drive wheel set includes a drive ball joint, an elastic support frame, a hub motor, and a drive wheel. The drive wheel is integrated on the outer rotor of the hub motor, and the hub motor is mounted on the elastic support frame. The top of the elastic support frame is connected to a drive ball socket on a fixed arc arm or a movable arc arm through a drive ball joint.

5. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 4, characterized in that, The surface of the drive wheel is a concave arc surface.

6. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, Both the first and second arc-shaped arms are equipped with pressurized water spraying components on their sidewalls near the direction of the first surrounding mechanism. The pressurized water spraying components include a pressurized water spraying pump and a spraying arm. The inlet end of the pressurized water spraying pump is equipped with a filter screen, and the outlet end of the pressurized water spraying pump is connected to a spraying arm with a hollow structure. The spraying arm is C-shaped and is equipped with several nozzles connected to the hollow structure.

7. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1 or 6, characterized in that, Several booster propellers are provided on the outer walls of both the first and second arc-shaped arms.

8. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, Both ends of the threaded rod are mounted in the mounting base via bearing seats.

9. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, The decoupled propulsion assembly includes four decoupled thrusters, which are symmetrically distributed in pairs, with an upper and lower position difference between each pair of decoupled thrusters.

10. The submarine cable fault detection robot based on a biomimetic inchworm structure according to claim 1, characterized in that, It also includes a control mechanism and a detection system. The control mechanism and detection system are equipped with a fault diagnosis neural network model constructed using a convolutional neural network and a multi-head self-attention mechanism. The information collected by the sensing component is input into the fault diagnosis neural network model for analysis, and the cable detection results are output.

Images