A Dedicated Deep-Sea Multi-Robot Autonomous Mining System and Method

By utilizing a stationary deep-sea multi-robot autonomous mining system, which leverages seabed relay stations and underwater robot swarms to work collaboratively, the environmental disturbance and efficiency issues in deep-sea metal nodule extraction have been resolved, achieving precise extraction and efficient operation.

CN123109666APending Publication Date: 2026-11-17SOUTHEAST UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611348420.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing methods for collecting deep-sea metal nodules suffer from problems such as significant environmental disturbance, low efficiency of single-robot operations, limited continuous operation capabilities, and insufficient ability to identify, locate, coordinate, and accurately collect metal nodules in complex seabed environments.

Method used

The system employs a resident deep-sea multi-robot autonomous mining system, including a seabed relay station and an underwater robot cluster. The relay station provides power replenishment, ore transfer and storage, and task scheduling. The underwater robots perform search, location, mining, and return-to-station operations. It utilizes horseshoe crab-inspired bio-inspired disturbance-resistant bodies and multiple robotic arms to work collaboratively, and combines vision, electromagnetic, and contact force sensing units to achieve precise data collection.

Benefits of technology

It enables underwater refueling, ore transfer and storage, task scheduling, multi-robot collaborative operation, and precise end-point collection, reducing disturbance to the seabed environment, improving mining efficiency and stability, and enhancing operational capabilities in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123109666A_ABST
    Figure CN123109666A_ABST
Patent Text Reader

Abstract

This invention discloses a stationary deep-sea multi-robot autonomous mining system and method, comprising a seabed relay station deployed in a deep-sea mining area and an underwater robot cluster operating around the seabed relay station. The seabed relay station, as a stationary operation node, provides the underwater robot cluster with underwater power replenishment, ore transfer and storage, task scheduling, and data interaction. The underwater robot cluster includes several underwater robots that, under the scheduling of the seabed relay station, perform metal nodule search, location, mining, ore return, and return-to-station power replenishment operations. This invention addresses the problems of significant environmental disturbance, low efficiency of single-robot operations, limited continuous operation capabilities, and insufficient ability to identify, locate, coordinate, and accurately collect metal nodules in complex seabed environments in existing deep-sea mining methods, achieving seabed stationing, multi-robot collaboration, ore transfer, and underwater power replenishment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of deep-sea resource exploration and development, and in particular to a resident deep-sea multi-robot autonomous mining system and method. Background Technology

[0002] Deep-sea metal nodule extraction is an important direction for deep-sea mineral resource development. Polymetallic nodules are widely distributed in the surface layer of deep-sea seabed sediments, rich in metallic elements such as manganese, nickel, cobalt, and copper, and have high development and utilization value. Compared with terrestrial mining, deep-sea metal nodule extraction faces special environmental constraints such as high pressure, low light, low temperature, limited communication, bottom current disturbance, and complex terrain. Extraction equipment not only needs to possess near-bottom stable movement, autonomous identification and positioning capabilities, and precise extraction capabilities, but also needs to minimize disturbance to seabed sediments and the benthic ecosystem. Therefore, developing technologies for low-disturbance, precise extraction, seabed refueling and transfer, robotic collaborative operations, and autonomous return-to-station cyclic operations for the near-bottom environment is of great significance for the green and efficient development of deep-sea metal nodules.

[0003] Current deep-sea metal nodule collection primarily relies on large-scale mobile ore collection vehicles, supplemented by hydraulic or pipeline lifting devices for ore recovery. While this method enables continuous collection, it easily disturbs seabed sediments and creates plumes, resulting in significant environmental impact. Using underwater robots with robotic arms for targeted collection can reduce the disturbance caused by traditional methods. However, existing single robots are limited by battery capacity, ore storage space, communication distance, and operating radius, making long-term seabed stays difficult. Furthermore, the collection coverage of a single robot or robotic arm is limited, leading to low efficiency when dealing with discretely distributed metal nodules. In conditions of near-bottom current disturbance, sediment cover, and complex terrain, problems such as inaccurate target identification, robotic arm positioning deviation, mis-grabbing or missed grabbing, unstable gripping, and low efficiency in returning to the station for recharging and unloading can easily arise. Summary of the Invention

[0004] Purpose of the invention: The purpose of this invention is to provide a stationary deep-sea multi-robot autonomous mining system and method to solve the problems of large environmental disturbances, low efficiency of single robot operation, limited continuous operation capability, and insufficient ability to identify, locate, coordinate and schedule and accurately collect metal nodules in complex seabed environments in existing deep-sea mining methods. It realizes seabed stationary operation, multi-robot collaboration, autonomous search and positioning, accurate collection, ore transfer and underwater refueling.

[0005] Technical Solution: The resident deep-sea multi-robot autonomous mining system of the present invention includes a seabed relay station deployed in a deep-sea mining area and an underwater robot cluster operating around the seabed relay station; the seabed relay station serves as a resident operation node, used to provide underwater power replenishment, ore transfer and storage, task scheduling, and data interaction for the underwater robot cluster; the underwater robot cluster includes several underwater operating robots, which are used to perform metal nodule search, location, mining, ore return, and return-to-station power replenishment operations under the scheduling of the seabed relay station; when the temporary ore stored in the ore storage bin of the underwater operating robot reaches the capacity threshold, or the remaining power is lower than the safe power threshold for returning to the station, the underwater operating robot returns to the seabed relay station for ore unloading and underwater power replenishment.

[0006] Furthermore, the subsea relay station includes a main frame, stationary support legs, a guiding docking and positioning structure, an ore transfer bin, and an ore docking guide port. The guiding docking and positioning structure includes an outwardly expanding guide surface, a converging guide rail, a limiting block, and a positioning detection component, and is located in the docking service area of ​​the subsea relay station to guide the underwater robot to the unloading and refueling position. The bottom of the ore storage bin facing the return docking side is provided with an unloading port and an openable gate. After docking, the unloading port and the ore docking guide port are vertically aligned, allowing the collected ore to enter the ore transfer bin through the ore docking guide port.

[0007] Furthermore, the seabed relay station also includes a data interaction module and a task scheduling module;

[0008] The data interaction module is used to establish a data communication link between the seabed relay station and the underwater operation robot, and to receive, cache and forward the underwater operation robot's operation data, remaining power, ore storage bin loading, location status, equipment status and docking status; the data interaction module is also used to send the operation instructions, return instructions, recharge instructions and docking instructions generated by the task scheduling module to the corresponding underwater operation robot.

[0009] The task scheduling module includes a processor, a memory, and a communication interface connected to the data interaction module. The memory stores a scheduling program that can be executed by the processor. The processor is used to generate the underwater operation robot's operation area allocation, task objectives, task priorities, number of dispatches, return-to-station order, recharging order, charging power, and departure schedule based on the underwater operation robot's status information obtained by the data interaction module, combined with mineral distribution density, operation area location, seabed topography information, relay station energy storage status, remaining capacity of the transfer ore bin, remaining robot power, ore temporary storage bin loading capacity, distance between the underwater operation robot and the seabed relay station, and other underwater operation robot locations.

[0010] Furthermore, the underwater operation robot includes a horseshoe crab-inspired bionic disturbance-resistant body and several independent mining robotic arm units arranged on the ventral side of the horseshoe crab-inspired bionic disturbance-resistant body.

[0011] Each independent mining robotic arm unit corresponds to a different main mining sector. The independent mining robotic arm unit is deployed during the operation period for local exposure, clamping and extraction of ore; during the non-operation period, it is retracted and locked along the ventral side of the horseshoe crab-like bionic anti-interference body.

[0012] Furthermore, the underwater robot also includes a multi-source sensing module, an autonomous control module, and a propulsion attitude adjustment unit;

[0013] The multi-source sensing module is used to identify ore targets, sense the near-bottom environment, and assist the robot in attitude stabilization, path planning, and docking with the seabed relay station.

[0014] The autonomous control module is used to control the propulsion attitude adjustment unit to complete navigation, hovering, attitude stabilization and return docking based on the information obtained by the multi-source sensing module, and to control the independent mining robotic arm unit to complete target approach, partial exposure, clamping and extraction, and transfer the collected ore to the ore temporary storage bin.

[0015] Furthermore, the propulsion attitude adjustment unit includes multiple underwater thrusters, pressure-resistant actuators, and mounting brackets distributed around the circumference of the horseshoe crab-inspired biomimetic anti-disturbance body. The multiple underwater thrusters generate longitudinal, lateral, and vertical thrusts, as well as roll, pitch, and yaw control torques through thrust vector combinations.

[0016] Furthermore, the independent mining robotic arm unit includes a multi-degree-of-freedom robotic arm body, an end-effector acquisition component disposed at the end of the multi-degree-of-freedom robotic arm body, and a closed-loop mining control unit; the end-effector acquisition component includes a shovel-clamp integrated end-effector acquisition device and an end-effector sensing component;

[0017] The multi-degree-of-freedom robotic arm body is used to drive the end-effector acquisition component to approach the ore target; the end-effector sensing component is used to acquire target position, mineral properties, contact state, clamping state and extraction state information; the closed-loop mining control unit is used to perform pre-mining confirmation, in-mining correction and post-mining verification based on the information acquired by the end-effector sensing component.

[0018] Furthermore, the end-sensing component includes a visual recognition unit, an electromagnetic confirmation unit, and a contact force sensing unit;

[0019] The visual recognition unit is used to identify the ore boundary, clamping point, and the relative position of the shovel-clamp integrated end collector and the ore.

[0020] The electromagnetic confirmation unit includes a detection coil, a reference coil, and a non-metallic isolation seat. The detection coil is located at the non-metallic isolation position at the front end of the integrated shovel and clamp end collector. The reference coil is located away from the ore contact area and close to the metal structure of the independent mining robotic arm unit. The electromagnetic confirmation unit compensates for the background response generated by the metal structure of the independent mining robotic arm unit by using the differential signal between the detection coil and the reference coil, combined with the pre-mining no-load background value, for pre-mining confirmation of the metal properties of suspected targets.

[0021] The contact force sensing unit is used to determine the contact state, clamping stability, slippage, and force changes during the grasping, scooping, and lifting processes.

[0022] Furthermore, in the pre-mining stage, the closed-loop mining control unit determines whether the suspected target is the ore to be mined based on the ore boundary, clamping point and relative position information obtained by the visual recognition unit and the metal attribute information obtained by the electromagnetic confirmation unit, and determines the approach path, clamping position and collection method of the shovel-clamp integrated end collector.

[0023] During the mining phase, the closed-loop mining control unit, based on the contact state, clamping force, sliding state, shovel entry resistance, and lifting force changes fed back by the contact force sensing unit, and in conjunction with the vision recognition unit, corrects in real time the grasping angle, clamping force, shovel entry depth, and lifting trajectory of the independent mining robotic arm unit.

[0024] In the post-mining stage, the closed-loop mining control unit determines whether the ore has been successfully mined and sent to the ore storage bin based on the images before and after mining acquired by the visual recognition unit, the post-mining signal from the electromagnetic confirmation unit, the force changes before and after unloading from the contact force sensing unit, and the ore entry and loading status. When it is determined that there is a mining failure, unstable clamping, ore slippage, or failure to enter the bin, the closed-loop mining control unit re-determines the clamping point, adjusts the posture of the shovel-clamp integrated end collector, optimizes the shovel-clamp action parameters, and controls the shovel-clamp integrated end collector to perform a second mining operation.

[0025] The underwater robot includes an ore storage bin; the ore storage bin is equipped with a weight detection unit, an entry detection unit, and an anti-blocking structure to determine the ore loading status and whether the ore is smoothly entering the bin.

[0026] Furthermore, the shovel-clamp integrated end collector includes a shovel front edge and a clamping part; for target ore exposed on the seabed surface, the clamping part directly clamps and extracts it; for partially buried target ore, the shovel front edge first slightly exposes and lifts the target ore, and then the clamping part completes the clamping and extraction.

[0027] The autonomous deep-sea multi-robot mining method described in this invention includes the following steps:

[0028] Deploy seabed relay stations and underwater robot clusters operating around the seabed relay stations in deep-sea mining areas;

[0029] The seabed relay station serves as a stationary operation node, providing underwater power replenishment, ore transfer and storage, task scheduling, and data interaction for the underwater robot cluster;

[0030] The underwater robot cluster includes several underwater operation robots, which perform metal nodule search, location, mining, ore return and refueling operations under the scheduling of the seabed relay station;

[0031] When the amount of ore stored in the ore storage bin of the underwater robot reaches the capacity threshold, or when the remaining power is lower than the safe power threshold for returning to the station, the underwater robot returns to the seabed relay station to unload the ore and replenish its power underwater.

[0032] Beneficial Effects: Compared with existing technologies, this invention simultaneously possesses the capabilities of seabed relocation and refueling, ore transfer and storage, task scheduling, data interaction, multi-robot collaborative operation, near-bottom disturbance-resistant stable motion, and precise end-point acquisition. Facing the deep-sea near-bottom operating environment, this invention constructs a seabed-based relocation relay station as a refueling, unloading, and scheduling node; configures multiple underwater robots to perform search, identification, positioning, acquisition, and return-to-station tasks around the relay station; employs a horseshoe crab-inspired biomimetic disturbance-resistant body to improve near-bottom stability; expands the acquisition range and reduces robot interference through the partitioned arrangement of multiple robotic arms on the ventral side; and integrates a shovel-gripper integrated acquisition device, as well as visual recognition, electromagnetic confirmation, and contact force sensing units at the end of the robotic arms, forming a closed-loop mining process of pre-mining confirmation, in-mining correction, post-mining verification, and secondary acquisition after failure. Specifically, this invention has the following advantages:

[0033] 1. This invention enables underwater power replenishment, ore transfer and storage, task scheduling and data exchange through a seabed relay station, reducing the time loss caused by the frequent surfacing and recovery of underwater robots and improving the continuous operation capability of deep-sea mining.

[0034] 2. This invention improves the near-bottom collection coverage of a single robot and the mining efficiency per unit time by coordinating the operation of an underwater robot cluster and multiple independent mining robotic arm units.

[0035] 3. This invention improves the stability, maneuverability, and operational safety of the underwater robot 200 in complex near-seabed environments by using a horseshoe crab-inspired bio-inspired anti-disturbance body and a ventral robotic arm folding structure.

[0036] 4. This invention achieves pre-mining confirmation, in-mining correction, and post-mining verification through a multi-source sensing closed loop of vision, electromagnetics, and contact force, thereby improving the success rate of metal nodule identification, positioning, clamping, and collection.

[0037] 5. The present invention adopts a fixed-point precise collection method, which can reduce the disturbance to seabed sediments caused by traditional large-scale ore collection and hydraulic methods, and is conducive to reducing the impact of the mining process on the seabed environment. Attached Figure Description

[0038] Figure 1 This is a schematic diagram illustrating the collaborative operation of multiple underwater robots around a seabed relay station according to an embodiment of the present invention.

[0039] Figure 2 This is a schematic diagram of the structure of a submarine relay station according to an embodiment of the present invention.

[0040] Figure 3 This is a schematic diagram of the underwater operation robot according to an embodiment of the present invention.

[0041] Figure 4 A schematic diagram of the underwater robot in its retracted state, as shown in this embodiment of the invention.

[0042] Figure 5 This is a three-dimensional structural diagram of an independent mining robotic arm unit according to an embodiment of the present invention.

[0043] Figure 6 for Figure 5 The enlarged view of area A shows the arrangement of the integrated shovel-clamp end-effector 350, the shovel leading edge 351, the clamping part 352, and the end-effector sensing components.

[0044] Figure 7 This is a side view of an independent mining robotic arm unit according to an embodiment of the present invention.

[0045] Figure 8 This is a schematic diagram of the ore conveying and anti-blocking structure according to an embodiment of the present invention.

[0046] Figure 9 This is a schematic diagram of the docking, ore unloading, and energy replenishment structure between the underwater operation robot and the seabed relay station, according to an embodiment of the present invention.

[0047] Figure 10 This is a flowchart illustrating the operation of an embodiment of the present invention.

[0048] Figure label:

[0049] 100. Submarine relay station; 110. Relay station main frame; 120. Top communication and control cabin; 130. Ore transfer bin; 140. Ore docking and guide port; 150. Underwater power supply module; 160. Data interaction module; 170. Stationary support leg; 180. Guiding docking and positioning structure; 181. Outwardly expanding guide surface; 182. Converging guide rail; 183. Limiting block; 184. Arrival detection component; 200. Underwater operation robot; 210. Horseshoe crab-inspired bionic anti-interference body; 220. Ore temporary storage bin; 221. Unloading port; 222. Openable and closable gate; 223. Inclined bin bottom; 224. Pushing component; 225. Weight detection unit; 230. Multi-source sensor module; 240. Independent mining robotic arm unit; 250. End-effector acquisition component; 260, Propulsion posture adjustment unit; 270, Robotic arm retraction locking part; 280, Docking and energy replenishment receiving part; 290, Ore conveying mechanism; 291, Receiving bucket; 292, Circular chain plate; 293, Driving sprocket; 294, Driven sprocket; 295, Tensioning component; 296, Side guard; 297, Inlet guide chute; 298, Inlet detection unit; 299, Vibrating component; 310, Robotic arm base; 320, Upper arm connecting rod; 330, Lower arm connecting rod; 340, End-effector connector; 350, Shovel-clamp integrated end-effector acquisition device; 351, Shovel leading edge; 352, Clamping part; 360, Visual recognition unit; 370, Electromagnetic confirmation unit; 380, Contact force sensing unit; 390, Drive joint module. Detailed Implementation

[0050] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0051] Example 1

[0052] like Figure 1 As shown, the resident deep-sea multi-robot autonomous mining system of the present invention includes a seabed relay station 100 deployed in a deep-sea mining area and an underwater robot cluster operating around the seabed relay station 100; the seabed relay station 100 serves as a resident operation node, used to provide underwater power replenishment, ore transfer and storage, task scheduling and data interaction for the underwater robot cluster; the underwater robot cluster includes several underwater operation robots 200, which are used to perform metal nodule search, location, mining, ore return and return-to-station power replenishment operations under the scheduling of the seabed relay station 100.

[0053] Specifically, multiple underwater robots 200 will conduct coordinated operations in designated areas around the seabed relay station 100. Figure 1Four underwater robots 200 are shown, each corresponding to work areas A through D. Red arrows indicate the departure paths of the underwater robots 200, while blue arrows indicate their return paths for unloading ore or recharging. A subsea relay station 100 is located between the work areas. Multiple underwater robots 200 enter their respective work areas to perform metal nodule search, location, and collection tasks, forming a cyclical operation relationship with the subsea relay station 100 through their departure and return paths for unloading and recharging. The subsea relay station 100 updates its tasks based on the remaining power of each underwater robot 200, the loading status of the ore storage bin 220, its current location, the completion status of the work area, and the docking service status, enabling the robot cluster to form a continuous autonomous operation mode of "zoned search and collection, return for unloading and recharging, and departure again."

[0054] like Figure 2 As shown, the seabed relay station 100 includes a main frame 110, a top communication and control cabin 120, an ore transfer bin 130, an ore docking and feeding port 140, an underwater power supply module 150, a data interaction module 160, a stationary support leg 170, and a guiding docking and positioning structure 180. The main frame 110 of the relay station houses various functional modules; the top communication control cabin 120 contains a task scheduling module for task scheduling and status management; the ore transfer bin 130 temporarily stores ore returned by the underwater robot 200; the ore docking guide port 140 receives ore unloaded by the underwater robot 200; the underwater power supply module 150 provides energy to the underwater robot 200; the data interaction module 160 facilitates the exchange of operational and task data between the underwater robot 200 and the seabed relay station 100; the stationing support leg 170 is fixedly connected to the underside of the main frame 110 to support the seabed relay station 100 in stable stationing on the seabed; the guiding docking and positioning structure 180 is located in the docking service area of ​​the seabed relay station 100 to assist the underwater robot 200 in completing its return docking. An energy storage module is located inside the main frame 110 to provide energy for the underwater power supply module 150 and other electrical equipment within the station.

[0055] In this embodiment, the ore is a metal nodule. The deployment location of the seabed relay station 100 can be determined comprehensively based on the distribution information of the metal nodules, the seabed topography information, the endurance radius of the underwater operation robot 200, the communication range, and the recovery conditions, so that the underwater operation robot 200 can maintain a reliable return-to-station capability while covering the target mining area.

[0056] This invention primarily targets metallic nodules for collection. The system architecture, including seabed relocation relay, multi-robot task scheduling, underwater refueling, data interaction, and autonomous return-to-station circulation, can also be applied to other deep-sea mineral resources such as cobalt-rich crusts and polymetallic sulfides. Depending on the ore body's occurrence state and mechanical properties, the shovel-clamp integrated end collector 350 in the end-collection component 250 can be replaced with a cutting, crushing, stripping, or suction collector, and the ore conveying mechanism and the feeding structure of the ore storage bin 220 can be adjusted accordingly.

[0057] In this embodiment, the seabed relay station 100 obtains the status information of the underwater operation robot 200 through the task scheduling module, and determines the dispatching robot, operation area, task objective, return order and refueling order based on the mineral distribution density, operation area location, relay station energy storage status, remaining capacity of transfer ore bin, remaining power of robot and loading capacity of ore temporary storage bin 220.

[0058] In this embodiment, the underwater power replenishment module 150 includes an underwater docking and positioning unit and a non-contact charging unit. After the underwater operation robot 200 returns to the seabed relay station 100, it completes docking and positioning through the underwater docking and positioning unit and replenishes its underwater power through the non-contact charging unit. The non-contact charging unit adopts an inductive coupling charging structure, which includes a transmitting coil disposed on the seabed relay station 100 and a receiving coil disposed on the underwater operation robot 200.

[0059] In this embodiment, the task scheduling module includes a processor, a memory, and a communication interface connected to the data interaction module 160. The memory stores a scheduling program that can be executed by the processor. The processor is used to generate the underwater operation robot 200's operation area allocation, task objectives, task priorities, number of dispatches, return-to-station order, recharging order, charging power, and departure-again-station arrangement based on the underwater operation robot 200's status information obtained by the data interaction module 160, combined with mineral distribution density, operation area location, seabed topography information, relay station energy storage status, remaining capacity of the transfer ore bin, remaining robot power, loading capacity of the ore temporary storage bin 220, distance between the underwater operation robot 200 and the seabed relay station 100, and the locations of other underwater operation robots 200.

[0060] In this embodiment, the data interaction module 160 is used to establish a data communication link between the seabed relay station 100 and the underwater operation robot 200, and to receive, cache and forward the operation data, remaining power, loading capacity of the ore storage bin 220, location status, equipment status and docking status of the underwater operation robot 200; the data interaction module 160 is also used to send the operation instructions, return instructions, recharge instructions and docking instructions generated by the task scheduling module to the corresponding underwater operation robot 200.

[0061] like Figure 3 As shown, the underwater operation robot 200 includes a horseshoe crab-inspired bionic disturbance-resistant body 210, an ore storage bin 220, a multi-source sensor module 230, an independent mining robotic arm unit 240 and its end effector acquisition component 250, a propulsion attitude adjustment unit 260, a robotic arm retraction locking part 270, a docking power receiving part 280, an autonomous control module, and an ore conveying mechanism. The robotic arm retraction locking part 270 consists of a fixed locking seat fixed to the ventral side of the body 210 and a matching locking component disposed on the robotic arm unit 240; the two engage when the robot is retracted into place. An independent mining robotic arm unit 240 is positioned on the ventral side of the horseshoe crab-inspired bionic anti-interference body 210. During non-operational periods, it is retracted and secured by the robotic arm retraction locking part 270. During operational periods, it extends to the seabed collection area. Its end-effector is equipped with an end-effector collection component 250, which includes a shovel-clamp integrated end-effector collector 350 and an end-effector sensing component. The shovel-clamp integrated end-effector collector 350 is used to collect seabed metal nodules and transfer the collected ore to an ore conveying mechanism, which then sends it into an ore storage bin 220. An autonomous control module is located inside the underwater robot 200. This module controls the robot's navigation, obstacle avoidance, robotic arm mining, return-to-station docking, and recharging processes based on information acquired by the multi-source sensor module 230. Specifically, this includes ore target identification, mineral property confirmation, near-bottom working height control, seabed terrain perception, obstacle avoidance, attitude stabilization, robotic arm mining positioning, and autonomous docking with external resident nodes.

[0062] In this embodiment, the operational period refers to the mining operation period, while the non-operational period includes the stages of departure patrol, mine area transfer, return to station for unloading, docking for refueling, and docking with external station nodes. During the non-operational period, the independent mining robotic arm unit 240 retracts and locks along the ventral side of the horseshoe crab-like bionic anti-interference body 210, and the robotic arm end-effector acquisition component 250 retracts into the ventral protection area within the projection range of the underwater robot 200's shell. The robotic arm retraction and locking part 270 includes a fixed locking seat fixed to the ventral side of the horseshoe crab-like bionic anti-interference body 210 and a cooperating locking component disposed on the independent mining robotic arm unit 240; such as Figure 4 As shown, when the robotic arm retracts into position, the fixed lock seat engages with the cooperating locking component to restrict the swinging and extension of the robotic arm relative to the body.

[0063] In this embodiment, the horseshoe crab-inspired bio-inspired disturbance-resistant body 210 adopts a wide, flat, streamlined shape and a low center of gravity structure to improve the robot's disturbance-resistant stability during near-bottom operations. The wide, flat, low-profile shell of the horseshoe crab-inspired bio-inspired disturbance-resistant body 210 can reduce the impact of water flow during near-seabed navigation and hovering, reduce yaw, pitch, and roll disturbances under the action of near-bottom water flow, and provide stable body support for the extension, gripping, and ore lifting of the robotic arm during mining operations.

[0064] In this embodiment, the multi-source sensing module 230 includes a visual sensor, an electromagnetic sensor, a pressure sensor, a near-bottom height sensor, a sonar, and an attitude sensor. The underwater robot 200 acquires near-bottom environment information, metal nodule target information, robot attitude information, and obstacle information through the multi-source sensing module 230, and autonomously completes target search, path planning, obstacle avoidance, and mining action control based on the above information.

[0065] In this embodiment, the propulsion attitude adjustment unit 260 includes multiple underwater thrusters, pressure-resistant actuators, and mounting brackets distributed around the circumference of the horseshoe crab-inspired biomimetic anti-interference body 210. The multiple underwater thrusters generate longitudinal, lateral, and vertical thrust, as well as roll, pitch, and yaw control torques through thrust vector combinations. The autonomous control module generates propulsion control commands based on the altitude, heading, attitude, and obstacle information output by the multi-source sensor module 230, and controls the propulsion attitude adjustment unit 260 to perform forward, backward, lateral, heave, near-bottom hovering, trajectory tracking, attitude stabilization, and low-speed fine-tuning during return docking. Therefore, the underwater robot 200 moves using a thruster-driven near-bottom swimming and hovering method.

[0066] In this embodiment, the underwater robot 200 uses a multi-source sensor module 230 to perceive the near-bottom environment, identify metal nodules, locate targets, and avoid obstacles. Based on seabed topography, target location, the loading status of the ore storage bin 220, remaining battery power, and distance to the seabed relay station 100, it autonomously plans search paths, mining paths, and return paths. The seabed relay station 100 allocates work areas and schedules work tasks, while the underwater robot 200 performs real-time near-bottom perception, path planning, and obstacle avoidance control.

[0067] In this embodiment, multiple independent mining robotic arm units 240 are arranged in sections around the ventral working area of ​​the underwater robot 200. Each independent mining robotic arm unit 240 corresponds to a preset collection sector. The multiple independent mining robotic arm units 240 coordinately execute collection tasks according to the target ore location and the occupancy status of adjacent collection sectors, enabling the robot to expand its collection coverage and improve mining efficiency per unit time in a near-bottom environment. The autonomous control module sets motion limits and interlock controls for the boundary areas between adjacent collection sectors based on the joint angles and end-effector positions of each independent mining robotic arm unit 240, thereby reducing the risk of cross-movement and mutual interference of the robotic arms.

[0068] In this embodiment, there are four independent mining robotic arm units 240, which are respectively arranged in the left front, right front, left rear and right rear regions on the ventral side of the horseshoe crab biomimetic anti-interference body 210 to form a four-arm zoned mining structure on the ventral side.

[0069] like Figure 5 , Figure 6 , Figure 7As shown, the independent mining robotic arm unit 240 includes a multi-degree-of-freedom robotic arm body, an end-effector acquisition component 250, and a closed-loop mining control unit. The end-effector acquisition component 250 includes a shovel-clamp integrated end-effector acquisition device 350 and an end-effector sensing component. The multi-degree-of-freedom robotic arm body includes a robotic arm base 310, a large arm link 320, a small arm link 330, an end-effector connector 340, and a drive joint module 390. The end-effector sensing component includes a vision recognition unit 360, an electromagnetic confirmation unit 370, and a contact force sensing unit 380. The robotic arm base 310 is mounted on the ventral side of the underwater robot 200. The robotic arm base 310 is rotatably connected to the large arm link 320, the large arm link 320 to the small arm link 330, and the small arm link 330 to the end-effector connector 340 through a drive joint module 390. The end-effector connector 340 is used to connect the end-effector acquisition component 250, and the end-effector acquisition component 250 carries the shovel-clamp integrated end-effector acquisition device 350 and the end-effector sensing component.

[0070] In this embodiment, the drive joint module 390 includes a base rotation joint, a shoulder pitch joint, a swing joint, a wrist posture adjustment joint, and an end effector connection structure, enabling the robotic arm to cover the lower front and abdominal working areas while the robot body is hovering. Each joint adopts a pressure-resistant and sealed drive structure, and is externally equipped with anti-corrosion materials and a structure to prevent mud and sand from entering, in order to adapt to the high pressure, low temperature, and sediment environment of the deep sea.

[0071] In this embodiment, the integrated shovel-clamp end-effector 350 includes a shovel-leading edge 351 and a clamping part 352. For metal nodules exposed on the seabed surface, the clamping part 352 directly clamps them; for metal nodules partially covered by sediment, the shovel-leading edge 351 first slightly exposes the target, and then the clamping part 352 clamps and extracts it. This structure concentrates the exposure, clamping, and extraction actions in the same end-effector, which helps to shorten the collection process and improve the collection success rate.

[0072] In this embodiment, the clamping part 352 is a gripper, and a flexible coating layer is also provided on the inner side of the gripper; for ores with irregular shape, slippery surface, and large size difference, the flexible coating layer is used to increase the contact area, improve the clamping stability and reduce the slippage of the ore.

[0073] In this embodiment, the visual recognition unit 360 is used to identify the ore boundary, gripping point, and the relative position between the collector and the ore. The electromagnetic confirmation unit 370 includes a detection coil, a reference coil, and a non-metallic isolation seat. The detection coil is located at the non-metallic isolation position at the bottom of the front end of the integrated shovel-grip end collector 350, and the reference coil is located away from the ore contact area and close to the metal structure of the robotic arm. Before collection, the controller collects the no-load background value, and during the mining process, it performs differential processing on the detection coil signal and the reference coil signal to cancel the common-mode background response generated by the changes in the metal structure and posture of the robotic arm, thereby improving the reliability of metal property confirmation for suspected metal nodules. The contact force sensing unit 380 is used to determine the contact state, gripping stability, slippage, and force changes during the gripping, shoveling, and lifting processes.

[0074] In this embodiment, the ore storage bin 220 is located in the middle of the underwater robot 200, and the ore conveying mechanism is located between the end-effector of the independent mining robotic arm unit 240 and the ore storage bin 220. After the independent mining robotic arm unit 240 transfers the collected ore to the ore conveying mechanism, the ore is then conveyed into the ore storage bin 220. The ore storage bin 220 may be equipped with a weight detection unit, an entry detection unit, and an anti-blocking structure to determine the loading status and whether the ore entry is smooth. When any of the following conditions occur: the ore storage bin 220 reaches a preset capacity threshold, or the robot's remaining battery power is lower than the safe return-to-station battery power threshold, the underwater robot 200 autonomously returns to the seabed relay station 100 for unloading ore and recharging. The loading status of the ore storage bin 220 is obtained by the weight detection unit, the ore entry status is obtained by the entry detection unit, the remaining battery power is obtained by the robot's battery management module, and the safe return-to-station battery power threshold is determined based on the return-to-station distance, return power consumption, and safety margin.

[0075] like Figure 8 As shown, the ore conveying mechanism 290 adopts a closed-type side-guard chain conveying structure, including a receiving hopper 291, annular chain plate 292, driving sprocket 293, driven sprocket 294, tensioning assembly 295, side guards 296, and inlet guide chute 297. The receiving hopper 291 is located within the unloading range of the end-collection assembly 250, and the inlet guide chute 297 is connected to the inlet of the ore temporary storage bin 220. After the shovel-clamp integrated end collector 350 releases the ore into the receiving hopper 291, the drive motor drives the driving sprocket 293 and annular chain plate 292 to send the ore into the inlet guide chute 297. A transverse baffle is provided on the annular chain plate 292, and a flexible sealing skirt is provided between the side baffle 296 and the annular chain plate 292 to limit the rolling of ore and reduce the entry of mud and sand; the tensioning component 295 is used to adjust the tension of the annular chain plate 292, and the annular chain plate 292 can run in the forward direction and run in the reverse direction for a short time.

[0076] A weight detection unit 225 is positioned between the ore storage bin 220 and the robot's support frame to obtain the ore loading amount based on the bin's load and determine whether the bin capacity threshold has been reached. An inlet detection unit 298 is positioned at the inlet guide chute 297 or the inlet of the ore storage bin 220 to detect the ore passage status. The anti-blocking structure includes an inclined guide surface, a vibrating element 299, and a ring chain plate 292 capable of forward and reverse operation. When the drive current or output torque exceeds the blockage threshold, the inlet detection signal duration exceeds a preset time, or ore is detected passing through but the weight detection unit 225 does not generate a corresponding weight increment, the autonomous control module determines that an inlet blockage has occurred and controls the ring chain plate 292 to briefly reverse and then forward, while simultaneously activating the vibrating element 299 to clear the blockage. If the blockage is not cleared within a preset number of cycles, the conveying is stopped, the fault is recorded, and a return-to-station process is triggered.

[0077] In this embodiment, pre-mining confirmation, in-mining correction, and post-mining verification are achieved through an end-sensing component. Specifically, in the pre-mining stage, the closed-loop mining control unit determines whether a suspected target is the ore to be mined based on the ore boundary, clamping point, and relative position information obtained by the visual recognition unit 360 and the metal attribute information obtained by the nearby electromagnetic confirmation unit 370, and determines the approach path, clamping position, and collection method of the integrated shovel-clamp end-sensor 350. In the in-mining stage, the closed-loop mining control unit, based on the contact state, clamping force, slippage state, shovel entry resistance, and lifting force changes fed back by the contact force sensing unit 380, and in conjunction with the visual recognition unit 360, corrects in real time the gripping angle, clamping force, shovel entry depth, and lifting trajectory of the independent mining robotic arm unit 240. In the post-mining stage, the closed-loop mining control unit determines whether the ore has been successfully mined and sent to the ore storage bin 220 based on the images before and after mining acquired by the visual recognition unit 360, the post-mining signal from the electromagnetic confirmation unit 370, the force changes before and after unloading from the contact force sensing unit 380, and the ore entry and loading status. When it is determined that there is a mining failure, unstable clamping, ore slippage, or failure to enter the bin, the closed-loop mining control unit re-determines the clamping point, adjusts the posture of the shovel-clamp integrated end collector 350 and optimizes the shovel-clamp action parameters, and controls the shovel-clamp integrated end collector 350 to perform a second mining, thereby realizing closed-loop control of search, mining, judgment, return to station, and interactive re-dispatch.

[0078] During post-mining verification, the visual recognition unit 360 compares images of the target area before and after acquisition to confirm whether the target has left its original position and whether the integrated shovel-clamp end collector 350 has completed release; the electromagnetic confirmation unit 370 determines whether the metal ore is still retained in the collector by judging whether the detection signal returns to the unloaded background range; the contact force sensing unit 380 determines whether the ore has been extracted and released based on the load change and the residual force after unloading; the closed-loop mining control unit then combines the pass signal from the bin entry detection unit and the weight increment from the weight detection unit to confirm whether the ore has been successfully entered into the bin. If the above signals are inconsistent, the closed-loop mining control unit performs secondary acquisition, re-release, or re-transmission.

[0079] Furthermore, such as Figure 9 As shown, after meeting the return-to-station conditions, the underwater operation robot 200 autonomously returns to the seabed relay station 100, completes positioning, ore unloading, energy replenishment, and data interaction, and then departs for another operation. When the underwater operation robot 200 returns to the station, the autonomous control module first controls the propulsion attitude adjustment unit 260 to complete coarse positioning based on acoustic positioning information and visual or sonar ranging information. Then, the robot's side guide enters the guiding docking positioning structure 180. The guiding docking positioning structure 180 includes an expanding guide surface 181, a converging guide rail 182, a limiting block 183, and a positioning detection component 184. The expanding guide surface 181 and the converging guide rail 182 gradually correct the robot's lateral and directional deviations, the limiting block 183 limits the longitudinal docking position, and the positioning detection component 184 outputs a final docking confirmation signal. The unloading end of the ore storage bin 220 is located at the bottom facing the return docking side, including an inclined bin bottom 223, an unloading port 221, an openable gate 222, and a pusher 224. At the final docking position, the unloading port 221 corresponds vertically to the ore docking guide port 140. After confirming the docking is in place, the openable gate 222 opens, and the collected ore, under the action of the inclined bin bottom 223 and the pusher 224, enters the ore transfer bin 130 through the ore docking guide port 140. Simultaneously, the docking energy receiving unit 280 corresponds to the underwater energy replenishment module 150 for underwater energy replenishment; the data interaction module 160 is used to confirm the return status, upload operation data, and exchange information for the next round of tasks. After completing unloading, energy replenishment, and data interaction, the underwater operation robot 200 leaves the station again to perform the next round of mining operations.

[0080] The working process of the resident deep-sea multi-robot autonomous mining system of the present invention is as follows: First, the deployment location of the seabed relay station 100 is determined according to the resource distribution of the target mining area, seabed topography, communication range, robot endurance radius and recovery conditions; then, the seabed relay station 100 allocates the working area to the underwater operation robot 200; after the underwater operation robot 200 leaves the station, it uses the multi-source sensing module 230 to perform near-bottom search, target recognition, path planning and obstacle avoidance; when a metal nodule target is identified, the independent mining robotic arm unit 240 approaches the target and completes pre-mining confirmation, in-mining correction and post-mining verification through the visual recognition unit 360, electromagnetic confirmation unit 370 and contact force sensing unit 380; the collected ore enters the ore temporary storage bin 220 through the ore conveying mechanism 290; when the return conditions are met, the underwater operation robot 200 returns to the seabed relay station 100, completes unloading, recharging and data interaction, and then leaves the station again to perform the next round of mining operations.

[0081] Example 2

[0082] like Figure 10 As shown, the autonomous deep-sea multi-robot mining method of the present invention includes the following steps.

[0083] Step 1: Based on the pre-obtained information on the distribution of metal nodules, seabed topography, the endurance radius of the underwater robot 200, its communication range, and recovery conditions, deploy relay stations in the deep-sea mining area.

[0084] Step 2: The relay station obtains the remaining power of each underwater operation robot 200, the current loading capacity of the ore storage bin 220, the location status, equipment status, docking status and other robot location status information through the data interaction module 160. The task scheduling module then determines the dispatching robot, the work area, the task objective, the task priority, the return order, the recharging order and the departure arrangement.

[0085] Step 3: After the underwater robot 200 leaves the relay station, it autonomously plans the search path, mining path and obstacle avoidance path based on the work area, task objectives and objectives assigned by the relay station, combined with the near-bottom environment information, metal nodule target location and obstacle information obtained by the multi-source sensor module 230.

[0086] Step 4: The underwater robot 200 controls the independent mining robotic arm unit 240 to approach the target metal nodule, and determines the target position, target attributes, gripping point and target exposure degree based on the ore boundary and relative position information obtained by the vision unit, the differential compensation signal between the detection coil and the reference coil in the electromagnetic unit, and the near-bottom environment information.

[0087] Step 5: The robotic arm forms a closed-loop mining process through pre-mining confirmation, in-mining correction and post-mining verification, and sends the collected ore into the ore temporary storage bin 220 via the ore conveying mechanism 290.

[0088] Step 6: When either the ore storage bin 220 reaches the storage capacity threshold or the remaining power is lower than the safe power threshold for returning to the station, the underwater robot 200 stops the current data collection task and autonomously plans its return path to return to the relay station.

[0089] Step 7: After completing ore unloading, underwater power replenishment, and data exchange at the relay station, the underwater operation robot 200 will receive task scheduling again and leave the station to execute the next round of mining operations.

Claims

1. A resident deep sea multi-robot autonomous mining system, characterized in that, The system includes a seabed relay station (100) deployed in a deep-sea mining area and an underwater robot cluster operating around the seabed relay station (100). The seabed relay station (100) serves as a resident operation node, providing underwater power replenishment, ore transfer and storage, task scheduling, and data interaction for the underwater robot cluster. The underwater robot cluster includes several underwater operation robots (200), which are used to perform metal nodule search, location, mining, ore return, and return-to-station power replenishment operations under the scheduling of the seabed relay station (100). When the temporary ore stored in the ore storage bin (220) of the underwater operation robot (200) reaches the capacity threshold, or the remaining power is lower than the safe power threshold for returning to the station, the underwater operation robot (200) returns to the seabed relay station (100) for ore unloading and underwater power replenishment.

2. The autonomous mining system of claim 1, wherein, The submarine relay station (100) includes a main frame (110), a stationing support leg (170), a guiding docking and positioning structure (180), an ore transfer bin (130), and an ore docking and feeding port (140). The stationing support leg (170) is fixedly connected to the bottom of the main frame (110) of the relay station and is used to support the submarine relay station (100) to stay on the seabed; The guiding docking and positioning structure (180) includes an outwardly expanding guide surface (181), a converging guide rail (182), a limiting block (183), and a positioning detection component (184), and is set in the docking service area of ​​the submarine relay station (100) to guide the underwater operation robot (200) to the unloading and refueling position; the bottom of the ore storage bin (220) facing the return docking side is provided with an unloading port (221) and an openable and closable gate (222). After docking, the unloading port (221) corresponds vertically to the ore docking guide port (140), so that the collected ore enters the ore transfer bin (130) through the ore docking guide port (140).

3. The autonomous mining system of claim 1, wherein, The submarine relay station (100) also includes a data interaction module (160) and a task scheduling module; The data interaction module (160) is used to establish a data communication link between the seabed relay station (100) and the underwater operation robot (200), and to receive, cache and forward the operation data, remaining power, loading capacity of the ore storage bin (220), location status, equipment status and docking status of the underwater operation robot (200); the data interaction module (160) is also used to send the operation instructions, return instructions, recharge instructions and docking instructions generated by the task scheduling module to the corresponding underwater operation robot (200). The task scheduling module includes a processor, a memory, and a communication interface connected to the data interaction module (160). The memory stores a scheduling program that can be executed by the processor. The processor is used to generate the underwater operation robot (200)'s operation area allocation, task objectives, task priorities, attendance quantity, return order, recharge order, charging power, and departure arrangement based on the underwater operation robot (200)'s status information obtained by the data interaction module (160), combined with mineral distribution density, operation area location, seabed topography information, relay station energy storage status, transfer ore bin remaining capacity, robot remaining power, ore temporary storage bin (220) loading capacity, distance between the underwater operation robot (200) and the seabed relay station (100), and other underwater operation robot (200) locations.

4. The autonomous mining system of claim 1, wherein, The underwater operation robot (200) includes a horseshoe crab bionic disturbance-resistant body (210) and several independent mining robotic arm units (240) arranged on the ventral side of the horseshoe crab bionic disturbance-resistant body (210). Each independent mining robotic arm unit (240) corresponds to a different main mining sector. The independent mining robotic arm unit (240) is deployed during the operation period to locally expose, clamp and extract the ore; during the non-operation period, it is retracted and locked along the ventral side of the horseshoe crab-like bionic anti-interference body (210).

5. The autonomous mining system of claim 1, wherein, The underwater robot (200) also includes a multi-source sensing module (230), an autonomous control module, and a propulsion attitude adjustment unit (260). The multi-source sensing module (230) is used to identify ore targets, sense the near-bottom environment and assist the robot in stabilizing its posture, planning its path and docking with the seabed relay station (100); The autonomous control module is used to control the propulsion attitude adjustment unit (260) to complete navigation, hovering, attitude stabilization and return docking based on the information obtained by the multi-source sensing module (230), and to control the independent mining robotic arm unit (240) to complete target approach, partial exposure, clamping and extraction, and transfer the collected ore to the ore temporary storage bin (220).

6. The autonomous mining system of claim 4, wherein, The independent mining robotic arm unit (240) includes a multi-degree-of-freedom robotic arm body, an end-effector acquisition component (250) disposed at the end of the multi-degree-of-freedom robotic arm body, and a closed-loop mining control unit; the end-effector acquisition component (250) includes a shovel-clamp integrated end-effector acquisition device (350) and an end-effector sensing component; The multi-degree-of-freedom robotic arm body is used to drive the end-effector acquisition component (250) to approach the ore target; the end-effector sensing component is used to acquire target position, mineral properties, contact state, clamping state and extraction state information; the closed-loop mining control unit is used to perform pre-mining confirmation, in-mining correction and post-mining verification based on the information acquired by the end-effector sensing component.

7. The autonomous mining system of claim 6, wherein, The end-sensing component includes a visual recognition unit (360), an electromagnetic confirmation unit (370), and a contact force sensing unit (380). The visual recognition unit (360) is used to identify the ore boundary, clamping point, and the relative position of the shovel-clamp integrated end collector (350) and the ore. The electromagnetic confirmation unit (370) includes a detection coil, a reference coil, and a non-metallic isolation seat. The detection coil is located at the non-metallic isolation position at the front end of the shovel-clamp integrated end collector (350). The reference coil is located away from the ore contact area and close to the metal structure of the independent mining robotic arm unit (240). The electromagnetic confirmation unit (370) compensates for the background response generated by the metal structure of the independent mining robotic arm unit (240) by using the differential signal between the detection coil and the reference coil, combined with the pre-mining no-load background value, for pre-mining confirmation of the metal properties of suspected targets. The contact force sensing unit (380) is used to determine the contact state, clamping stability, slippage, and force changes during the grasping, scooping, and lifting processes.

8. The autonomous mining system of claim 7, wherein, In the pre-mining stage, the closed-loop mining control unit determines whether the suspected target is the ore to be mined based on the ore boundary, clamping point and relative position information obtained by the visual recognition unit (360) and the metal attribute information obtained by the electromagnetic confirmation unit (370), and determines the approach path, clamping position and collection method of the shovel-clamp integrated end collector (350). During the mining phase, the closed-loop mining control unit, based on the contact state, clamping force, sliding state, shovel entry resistance and lifting force changes fed back by the contact force sensing unit (380), and in conjunction with the vision recognition unit (360), corrects in real time the gripping angle, clamping force, shovel entry depth and lifting trajectory of the independent mining robotic arm unit (240). In the post-mining stage, the closed-loop mining control unit determines whether the ore has been successfully mined and sent into the ore storage bin (220) based on the images before and after collection by the visual recognition unit (360), the post-mining signal of the electromagnetic confirmation unit (370), the force change before and after unloading by the contact force sensing unit (380), and the entry detection signal and weight increment of the ore storage bin (220). When it is determined that there is a collection failure, unstable clamping, ore slippage, or failure to enter the bin, the closed-loop mining control unit re-determines the clamping point, adjusts the posture of the shovel-clamp integrated end collector (350) and optimizes the shovel clamping action parameters, and controls the shovel-clamp integrated end collector (350) to perform secondary collection. The underwater operation robot (200) includes an ore storage bin (220); the ore storage bin (220) is equipped with a weight detection unit, an entry detection unit and an anti-blocking structure, which are used to determine the ore loading status and whether the ore enters the bin smoothly.

9. The autonomous mining system of claim 6, wherein, The integrated shovel and clamp end collector (350) includes a shovel front edge (351) and a clamping part (352); for target ore exposed on the seabed surface, the clamping part (352) directly clamps and extracts it; for partially buried target ore, the shovel front edge (351) first slightly exposes and lifts the target ore, and then the clamping part (352) completes the clamping and extraction.

10. A method for autonomous mining of deep-sea resources by a resident multi-robot system according to any one of claims 1 to 9, characterized in that, Includes the following steps: Deploy a seabed relay station (100) and an underwater robot cluster operating around the seabed relay station (100) in the deep-sea mining area; The seabed relay station (100) serves as a stationary operation node, providing underwater power replenishment, ore transfer and storage, task scheduling and data interaction for the underwater robot cluster; The underwater robot cluster includes several underwater operation robots (200), which perform metal nodule search, location, mining, ore return and refueling operations under the scheduling of the seabed relay station (100); When the amount of ore stored in the ore storage bin (220) of the underwater operation robot (200) reaches the capacity threshold, or when the remaining power is lower than the safe power threshold for returning to the station, the underwater operation robot (200) returns to the seabed relay station (100) to unload the ore and replenish its power underwater.