An autonomous positioning method and system for an underwater robot

Abstract

The application discloses an autonomous positioning method and system of an underwater robot, and belongs to the technical field of underwater robot positioning and navigation. At least one environmental parameter is collected according to a preset sensing frequency, and multi-source sensing data is collected, the multi-source sensing data being collected by a visual sensor, a sonar sensor, an inertial navigation unit and a force feedback sensor of a mechanical arm. According to the at least one environmental parameter collected for the first time, a weight distribution strategy is determined, and according to the weight distribution strategy and the multi-source sensing data, initial pose information is determined. Then, target pose information is determined based on touch force data collected by the force feedback sensor and the initial pose information, multi-sensor fusion positioning that is adaptive to the environment is realized, and pose correction is carried out based on the force feedback of the mechanical arm, so that the cumulative error divergence of the inertial navigation unit is effectively inhibited, the positioning error of the underwater robot during continuous operation of 72 hours can be controlled within 3 cm, and the positioning accuracy is improved by more than 50%.

Description

Technical Field

[0001] This application belongs to the field of underwater robot positioning and navigation technology, specifically relating to an autonomous positioning method, system, electronic device and storage medium for an underwater robot. Background Technology

[0002] With the rapid development of offshore wind power, marine resource exploration and other fields, underwater robots are increasingly being used in inspection and maintenance operations. However, existing underwater robot positioning technologies still have many technical bottlenecks, which seriously restrict their operational efficiency in complex marine environments.

[0003] Current underwater robot positioning technologies can be mainly categorized as follows: Vision-based positioning systems—these systems achieve positioning through optical cameras and target recognition. However, their positioning accuracy drops significantly in deep water areas with turbidity or poor lighting conditions. Furthermore, vision systems require high computational resources, making real-time performance difficult to guarantee. Sonar-based positioning systems—these systems use underwater acoustic beacons and relay buoy matrices for positioning. While suitable for long-distance transmission, they suffer from problems such as large signal delays (typically exceeding 200ms) and multipath effects, failing to meet the real-time positioning requirements in highly dynamic environments. Mechanical positioning systems—these systems achieve positioning control through mechanical structures such as turbines and rotating supports. Although they offer good stability, they lack flexibility and are prone to cumulative errors in turbulent water areas. Summary of the Invention

[0004] The purpose of this application is to provide an autonomous positioning method and system for underwater robots, which can solve the problems of poor environmental adaptability of single sensors, large signal delay, and uncontrollable cumulative error in existing underwater robot positioning technologies.

[0005] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, embodiments of this application provide an autonomous localization method for an underwater robot, applicable to an underwater robot, the method comprising: In response to underwater work instructions, at least one environmental parameter of the underwater environment is collected according to a preset sensing frequency; Collect multi-source sensor data during underwater operation; wherein, the multi-source sensor data is collected by multi-source sensors installed on the underwater robot, and the multi-source sensors include at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor of the robotic arm; Based on at least one environmental parameter acquired initially, a weight allocation strategy for the visual sensor, the sonar sensor, and the inertial navigation unit is determined, and initial pose information is determined based on the weight allocation strategy and the multi-source sensor data. Based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information, the target pose information is determined.

[0006] Optionally, the multi-source sensing data includes initial pixel images and initial sonar images acquired by the visual sensor and the sonar sensor at a first frequency, and inertial navigation data acquired by the inertial navigation unit at a second frequency. The step of determining the initial pose information based on the weight allocation strategy and the multi-source sensing data includes: Based on a preset filtering mechanism, feature point filtering processing is performed on the initial pixel image and the initial sonar image to obtain the target pixel image and the target sonar image. According to the weight allocation strategy, the target pixel map and the target sonar map at the first frequency and the inertial navigation data at the second frequency are fused together, and the initial pose information is output according to the preset output frequency.

[0007] Optionally, based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information, the target pose information is determined, including: Determine the confidence level of the initial pose information; If the confidence level of the initial pose information is greater than or equal to a preset working threshold, the initial pose information is used as the target pose information. If the confidence level of the initial pose information is lower than the preset working threshold, the initial pose information is corrected based on the tactile force data collected by the force feedback sensor in the multi-source sensing data to obtain the target pose information.

[0008] Optionally, the contact force data is collected by the force feedback sensor when the robotic arm contacts a spatial reference object in the underwater environment. The step of correcting the initial pose information based on the contact force data collected by the force feedback sensor to obtain the target pose information includes: Obtain prior information about the spatial reference object; Based on the contact force data and the prior information, the initial pose information is corrected to obtain the target pose information.

[0009] Optionally, it also includes: The weight allocation strategy is adjusted based on at least one currently collected environmental parameter and a preset adjustment range.

[0010] Optionally, it also includes: Fault detection is performed on the multi-source sensors; If any sensor malfunction is detected, data collection using the malfunctioning sensor is stopped, and the weight allocation strategy is redefined based on the non-malfunctioning sensors.

[0011] Optionally, it also includes: The confidence level of the target pose information is determined, and a relocation process is triggered if the confidence level of the target pose information is lower than the preset working threshold.

[0012] Optionally, the relocation process includes: Obtain the pre-stored prior map of the job scenario; The sonar sensor is controlled to acquire a relocation sonar map of the underwater environment, and the relocation sonar map is matched with a prior map of the operation scenario to obtain a matching result; Based on the matching result, the current pose information is determined, and the target pose information is updated using the current pose information.

[0013] Optionally, the underwater robot includes multiple preset working modes for different scenarios. The multiple preset working modes include an inspection mode, a precision operation mode, and an emergency mode. The inspection mode, the precision operation mode, and the emergency mode each have corresponding weight allocation strategies and preset output frequencies.

[0014] Secondly, embodiments of this application provide an autonomous positioning system for an underwater robot, applied to an underwater robot, the system comprising: The environmental sensing module is used to respond to underwater working instructions and collect at least one environmental parameter of the underwater environment according to a preset sensing frequency. A multi-source sensor module is used to collect multi-source sensing data during underwater operation; wherein, the multi-source sensing data is collected by multi-source sensors installed on the underwater robot, and the multi-source sensors include at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor of the robotic arm; The fusion positioning module is used to determine the weight allocation strategy of the visual sensor, the sonar sensor and the inertial navigation unit based on at least one environmental parameter acquired for the first time, and to determine the initial pose information based on the weight allocation strategy and the multi-source sensor data. The correction module is used to determine the target pose information based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information.

[0015] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0016] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0017] Fifthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0018] In this embodiment, in response to an underwater working command, at least one environmental parameter of the underwater environment is collected at a preset sensing frequency. Multi-source sensor data is collected during the underwater working process. This multi-source sensor data is collected by multi-source sensors installed on the underwater robot, including at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor from the robotic arm. Based on the first collected environmental parameter, a weight allocation strategy is determined for the vision sensor, the sonar sensor, and the inertial navigation unit. The weight allocation strategy and the multi-source sensor data are then used to determine the optimal data allocation strategy. According to the method, the initial pose information is determined, and the target pose information is determined based on the force feedback sensor collected by the force feedback sensor in the multi-source sensing data and the initial pose information. This realizes environmental adaptive multi-sensor fusion positioning. On this basis, the initial pose information is corrected based on the force feedback sensor collected by the force feedback sensor. The contact force vector of the robotic arm is used as the absolute correction source, which effectively suppresses the cumulative error divergence of the inertial navigation unit. It can keep the positioning error drift of the underwater robot within 3cm for 72 hours of continuous operation, which is more than 50% more accurate than the positioning error of more than 10cm in the existing solution.

[0019] Furthermore, it achieves dynamic determination of weight allocation strategy through environmental parameters, breaking through the environmental limitations of single sensor schemes and fixed weight modes. It can significantly improve positioning stability under harsh conditions such as turbid waters and strong current areas, avoid data conflicts and positioning jump problems, reduce the need for human intervention, and enhance the autonomous operation capability of underwater robots. Attached Figure Description

[0020] To more clearly illustrate the technical solution of this application, the drawings used in the description of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of the steps of an autonomous positioning method for an underwater robot provided in some embodiments of this application; Figure 2 This is a structural block diagram of an autonomous positioning system for an underwater robot provided in some embodiments of this application; Figure 3 This is a schematic diagram of the hardware structure of an electronic device provided in some embodiments of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] In existing technologies, such as the underwater ROV (Remotely Operated Vehicle) operation system disclosed in patent CN118760216A, although it integrates visual sensors, sonar, inertial navigation units and a robotic arm with force feedback, it only uses the force feedback of the robotic arm for trajectory adjustment and does not use it for absolute pose correction of the positioning system. This cannot solve the problem of cumulative error divergence in long-term operation of the inertial navigation unit. Moreover, its multi-sensor fusion adopts a fixed weight allocation strategy, which cannot adapt to complex and changeable marine conditions such as switching between clear and turbid waters and strong current disturbances. The positioning accuracy decreases significantly with environmental deterioration.

[0024] The underwater positioning system disclosed in patent CN120949243A can only adjust the weight of the inertial navigation unit in a single dimension according to the intensity of the robot's movement. It has not established a quantitative mapping relationship between marine environmental parameters (such as visibility, turbidity, and water flow speed) and the fusion weights of multiple sensors such as visual sensors, sonar, and inertial navigation units. Therefore, it cannot achieve environmental adaptive fusion in all scenarios and is prone to problems such as data conflict and positioning jump when operating across different working conditions, resulting in insufficient environmental adaptability.

[0025] The SLAM (Simultaneous Localization and Mapping) positioning method disclosed in patent CN116659510A does not perform lightweight adaptation for underwater embedded edge computing devices. It relies on the host computer on the water surface to complete the pose calculation, resulting in large transmission delays and failing to meet the needs of local real-time control. Furthermore, it does not incorporate prior information from underwater operation scenarios for optimization, leading to a high mismatch rate of feature points in the low-feature underwater environment and poor positioning robustness.

[0026] In summary, existing technologies generally suffer from the following key defects: single-sensor solutions have poor environmental adaptability; multi-sensor fusion lacks an intelligent weight allocation mechanism that adapts to the environment; pose correction can only achieve relative optimization and lacks a closed-loop correction mechanism based on an absolute reference, resulting in uncontrollable cumulative errors over long-term operation.

[0027] To address the aforementioned issues, this application utilizes multi-source sensor data fusion, a dynamic weight allocation strategy based on environmental parameters, and contact-based absolute pose correction based on a robotic arm force feedback sensor to achieve high-precision autonomous positioning of underwater robots in complex marine environments, effectively solving the problem of cumulative error divergence in inertial navigation units during long-term operations.

[0028] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0029] The following description, in conjunction with the accompanying drawings, details a method for autonomous positioning of an underwater robot provided in this application through specific embodiments and application scenarios.

[0030] Reference Figure 1 This document illustrates a flowchart of the autonomous localization method for an underwater robot according to some embodiments of this application. Applied to an underwater robot, the method may specifically include the following steps: Step 101: In response to the underwater work command, collect at least one environmental parameter of the underwater environment at a preset sensing frequency.

[0031] In practical applications, this method can be applied to underwater robots through an integrated autonomous positioning system. Specifically, the overall architecture of this autonomous positioning system can adopt a modular design, consisting of hardware subsystems, software subsystems, and auxiliary interfaces to ensure seamless integration with the ROV platform. Integration with the ROV platform can be achieved through standard interfaces such as RS485 (Recommended Standard 485) and Ethernet, allowing for easy access to mainstream ROV platforms and adaptability for upgrades to existing underwater equipment. The underwater robot described in this application is adaptable to a working temperature range of -20℃ to 50℃ and features a pressure-resistant design for a water depth of 300m, meeting the application requirements for most near-shore sea areas corresponding to the operational region.

[0032] I. The hardware subsystem includes a multi-source sensor module, a central processing unit (CPU), an umbilical cable transmission module, a power supply module, and an environmental sensing module.

[0033] Multi-source sensor module: This module includes multiple sensors, at least a high-definition optical camera (i.e., a visual sensor for visual positioning, supporting 4K resolution and optimized for underwater low-light conditions), a multi-beam imaging sonar (i.e., a sonar sensor for underwater acoustic imaging and environmental feature extraction), a high-precision inertial navigation unit (i.e., an inertial measurement unit (IMU) combined with a fiber optic gyroscope, with a sampling frequency ≥100Hz), and a six-dimensional force feedback sensor for the robotic arm (i.e., a force feedback sensor for the robotic arm, integrated into the ROV arm end for contact-based pose correction). The multiple sensors connect to the central processing unit via standardized interfaces, supporting "plug-and-play" and automatic calibration. After connection, the autonomous positioning system automatically completes spatiotemporal synchronization and calibration without manual intervention, facilitating rapid replacement and on-site maintenance.

[0034] Central Processing Unit (CPU): An embedded processor based on edge computing, responsible for real-time data computation, with computing power adapted to the requirements of lightweight SLAM algorithm operation. The CPU is connected to the surface control station via an umbilical cable to achieve uplink data transmission and downlink command transmission, and the core positioning calculations do not need to rely on surface equipment.

[0035] Umbilical cable transmission module: Serving as the power and communication backbone of the ROV, it provides a stable power supply and gigabit high-speed data link. The umbilical cable transmission module integrates a dual-redundancy design, including a primary communication link and a wireless acoustic backup link. Seamless switching between primary and backup links occurs within 10ms with no data packet loss, ensuring high reliability of data transmission.

[0036] Power module: This is a built-in intelligent power management system that supports power generation from the umbilical cable and is equipped with a small backup battery for emergency cable breakage.

[0037] Environmental sensing module: including temperature sensor, pressure sensor, flow rate and turbidity meter, etc., integrated into the ROV body, used to monitor marine conditions in real time.

[0038] II. The software subsystem includes a fusion positioning module, an intelligent control module, and a user interface module.

[0039] Fusion positioning module: Development environment-operating condition dual-dimensional fusion algorithm framework, supporting dynamic weight allocation, smooth transition of weight adjustment and fault weight reconstruction.

[0040] Correction Module: Integrates lightweight SLAM algorithm and robotic arm force feedback pose correction module.

[0041] Intelligent control module: This includes a fault detection module, an anomaly location and linkage path planning engine, and a full-process self-diagnostic tool. The module supports remote updates via firmware transmission over an umbilical cable.

[0042] Third, the auxiliary interface includes a user interface module, which provides real-time positioning visualization, position reliability assessment, fault alarm and full log recording for the surface control station software, and can be deeply integrated with the ROV control platform.

[0043] In step 101, firstly, the ROV obtains power and underwater operation commands from the surface control station via the umbilical cable. The autonomous positioning system initiates a full-process self-check, completes automatic sensor calibration, including visual focusing, sonar baseline, inertial navigation zero bias and spatiotemporal synchronization calibration, and loads a preset working mode according to the task.

[0044] Then, in response to the underwater operation command, the environmental sensing module is activated. Using temperature sensors, pressure sensors, flow meters, and turbidity meters, at a preset sensing frequency, at least one environmental parameter of the underwater environment is collected in real time and continuously. The environmental parameters include at least the water visibility, turbidity, water flow velocity, depth, temperature, and pressure of the underwater environment in which the underwater robot is located.

[0045] In some embodiments of this application, the underwater robot includes multiple working modes for different scenarios. The multiple working modes include an inspection mode, a precision operation mode, and an emergency mode. The inspection mode, the precision operation mode, and the emergency mode each have corresponding weight allocation strategies and preset output frequencies.

[0046] Specifically, underwater robots include multiple working modes for different scenarios, such as preset working modes for three core scenarios: offshore wind farm operation and maintenance, subsea pipeline inspection, and deep-sea equipment operation, with clearly defined parameter configurations for each preset working mode: (1) Inspection mode: Prioritizes battery life and large-area continuous positioning. The weight allocation strategy favors sonar and inertial navigation. The calculation frequency is 10Hz (i.e., the preset output frequency), and the overall power consumption is reduced by 30%. (2) Fine operation mode: Prioritize positioning accuracy, the weight allocation strategy favors vision and sonar, and closed-loop correction is performed simultaneously through the force feedback sensor of the robotic arm force. Positioning accuracy ≤ 5cm, calculation frequency 20Hz; (3) Emergency mode: Prioritize the survival and positioning capabilities of the autonomous positioning system, shut down unnecessary sensors, retain only the sonar sensor and inertial navigation unit, and enable low power operation strategy, reducing power consumption by 80%.

[0047] Step 102: Collect multi-source sensor data during underwater operation; wherein, the multi-source sensor data is collected by multi-source sensors installed on the underwater robot, and the multi-source sensors include at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor of the robotic arm.

[0048] In step 102, as learned in step 101, the multi-source sensor module includes at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor for the robotic arm. Specifically, the multi-source sensors continuously and synchronously acquire data in real time: the vision sensor identifies targets or environmental feature points, the sonar sensor scans the structure and features of the underwater environment, the inertial navigation unit records the underwater robot's posture and motion changes at high frequency, and the force feedback sensor for the robotic arm provides contact force data, which may include contact force vectors and contact position information, thereby obtaining multi-source sensor data of the underwater robot during its underwater operation.

[0049] Step 103: Based on at least one environmental parameter acquired in the first acquisition, determine the weight allocation strategy of the visual sensor, the sonar sensor, and the inertial navigation unit, and determine the initial pose information based on the weight allocation strategy and the multi-source sensor data.

[0050] In step 103, after collecting at least one environmental parameter of the underwater environment in real time and continuously using a temperature sensor, pressure sensor, flow meter, and turbidity meter at a preset sensing frequency, a weight allocation strategy for the visual sensor, the sonar sensor, and the inertial navigation unit can be determined based on the first collected environmental parameter.

[0051] Specifically, the underwater robot can quantify the current working conditions of the underwater environment into one of three typical marine working conditions based on the environmental parameters collected for the first time (such as visibility, turbidity, and water flow velocity): clear water low flow condition (visibility ≥ 5m, flow velocity ≤ 0.5m / s), turbid low flow condition (visibility < 5m, flow velocity ≤ 0.5m / s), or strong current disturbance condition (flow velocity > 0.5m / s, no visibility limit).

[0052] Based on the current operating conditions, a differentiated weight allocation strategy for the visual sensor, sonar sensor, and inertial navigation unit is determined: In clear water with low flow, the visual sensor accounts for 60% of the weight, sonar 25%, and inertial navigation 15%, achieving accurate positioning with high-resolution vision; in turbid water with low flow, the sonar sensor accounts for 60%, visual sensor 20%, and inertial navigation 20%, avoiding interference from turbidity in visual positioning; in strong current disturbance, the inertial navigation system accounts for 50%, sonar 35%, and visual sensor 15%, suppressing pose fluctuations caused by water flow disturbances through high-frequency inertial navigation data. After the weight allocation strategy is determined, the autonomous positioning system can fuse multi-source sensor data according to the weight allocation strategy to calculate the initial pose information.

[0053] In some embodiments of this application, it also includes: The weight allocation strategy is adjusted based on at least one currently collected environmental parameter and a preset adjustment range.

[0054] The autonomous positioning system continuously monitors environmental changes. The environmental perception module collects environmental parameters such as water visibility, turbidity, and water flow velocity in real time according to a preset sensing frequency. When environmental parameters change (e.g., the robot moves from clear water to turbid water, or encounters a sudden strong current disturbance), the autonomous positioning system dynamically adjusts the predetermined weight allocation strategy based on the currently collected environmental parameters to adapt to the changing operating conditions. In one example, the preset sensing frequency can be set to 10Hz.

[0055] Specifically, the autonomous positioning system employs a weighted smooth transition algorithm for weight adjustment. The update frequency of the weight allocation strategy is synchronized with the sampling frequency of environmental parameters (i.e., the preset sensing frequency). That is, after each new environmental parameter is collected, the autonomous positioning system determines the current public conditions and recalculates the target weight ratios of the visual sensor, sonar sensor, and inertial navigation unit based on the current operating conditions. During the adjustment process, the autonomous positioning system gradually corrects the weights in the weight allocation strategy according to a preset adjustment range, with a single round of weight adjustment not exceeding 10%, avoiding weight jumps caused by changes in operating conditions, thereby preventing sudden changes or fluctuations in positioning results. Thus, through multiple rounds of adjustments, the weight allocation strategy can smoothly transition to a weight ratio adapted to the new operating conditions. For example, from 60% vision / 25% sonar / 15% in clear water with low flow, it can be gradually adjusted to 60% sonar / 20% vision / 20% in turbid water with low flow.

[0056] Through the aforementioned dynamic adjustment mechanism, the autonomous positioning system can respond to changes in marine conditions in real time, achieve environmental adaptive fusion across all scenarios, and avoid data conflicts and positioning jumps that occur when operating across different conditions using the traditional fixed-weight mode.

[0057] In some embodiments of this application, the multi-source sensing data includes an initial pixel image and an initial sonar image acquired by the visual sensor and the sonar sensor at a first frequency, and inertial navigation data acquired by the inertial navigation unit at a second frequency. The step of determining the initial pose information based on the weight allocation strategy and the multi-source sensing data includes: Sub-step 11: Based on a preset filtering mechanism, feature point filtering processing is performed on the initial pixel image and the initial sonar image to obtain the target pixel image and the target sonar image.

[0058] In sub-step 11, the autonomous positioning system pre-screens feature points in the initial pixel image acquired by the visual sensor and the initial sonar image acquired by the sonar sensor, targeting the low-feature underwater environment. Specifically, low-confidence optical and sonar feature points can be eliminated based on two indicators: gradient threshold and neighborhood similarity. The gradient threshold is used to determine the drastic changes in brightness around a pixel; points with gradual changes have low gradient values ​​and are eliminated. Neighborhood similarity is used to determine whether the grayscale values ​​of a pixel are similar to those of its surrounding pixels; if the difference is too small, the point lacks distinctiveness and is unsuitable as a positioning feature point, and is also eliminated. After this screening process, high-confidence target pixel and target sonar images are obtained, reducing computational power consumption by more than 40%, thus ensuring real-time operation on the local embedded processor.

[0059] Sub-step 12: According to the weight allocation strategy, the target pixel map of the first frequency and the target sonar map, as well as the inertial navigation data of the second frequency, are fused together, and the initial pose information is output according to the preset output frequency.

[0060] In sub-step 12, the autonomous localization system employs a tightly coupled fusion framework. A lightweight SLAM algorithm fuses the high-frequency inertial navigation data acquired by the inertial navigation unit at a second frequency with the low-frequency target pixel image and target sonar image acquired by the visual and sonar sensors at a first frequency, after feature point filtering. The entire fusion calculation process is completed on the underwater robot's embedded central processing unit, without relying on surface equipment or relay buoys. Finally, it outputs the initial pose information according to a preset output frequency, with end-to-end latency controllable to within 20ms, meeting the real-time motion control requirements of the underwater robot.

[0061] In one example, the first frequency can be 10Hz, the second frequency can be 100Hz, and the preset output frequency can be 20Hz.

[0062] Step 104: Determine the target pose information based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information.

[0063] In step 104, the underwater robot comprehensively judges and determines the final target pose information based on the contact force data collected by the force feedback sensor and the initial pose information.

[0064] Specifically, the autonomous positioning system includes a pre-set real-time positioning accuracy evaluation model, which is used to evaluate the reliability of the initial pose information, and then decide whether to directly use the initial pose information or trigger absolute pose correction based on tactile force data, depending on the reliability of the initial pose information.

[0065] After the target pose information is determined, the target pose information can be simultaneously transmitted to the surface equipment via the umbilical cable. At the same time, the local central processing unit outputs the positioning results in real time, supporting ROV autonomous operation and fine-tuning without waiting for surface commands.

[0066] In some embodiments of this application, determining target pose information based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information includes: Sub-step 21: Determine the confidence level of the initial pose information.

[0067] In sub-step 21, the autonomous positioning system can evaluate the model in real time based on positioning accuracy. It calculates the confidence level of the initial pose information in real time based on the confidence level of multi-source sensor data, environmental parameters, and the frequency of corrections already performed. The confidence level reflects the reliability of the initial pose information; for example, the confidence level is high when the visual sensor collects abundant feature points in a clear water environment, and low when the water is turbid or the inertial navigation system has a large cumulative error. The autonomous positioning system can output the confidence interval of the positioning error in real time, providing a basis for subsequent decision-making.

[0068] Sub-step 22: If the confidence level of the initial pose information is greater than or equal to a preset working threshold, the initial pose information is used as the target pose information.

[0069] In sub-step 22, the autonomous positioning system can compare the confidence level of the initial pose information with a preset working threshold. The preset working threshold is a confidence level pre-set according to the requirements of the operation scenario, such as the confidence level corresponding to a 5cm positioning accuracy requirement.

[0070] If the confidence level of the initial pose information is greater than or equal to the preset working threshold, it indicates that the current positioning result is reliable and no additional correction is required. The autonomous positioning system directly outputs the initial pose information as the final target pose information for the motion control and operation execution of the underwater robot.

[0071] Sub-step 23: If the confidence level of the initial pose information is lower than the preset working threshold, the initial pose information is corrected based on the tactile force data collected by the force feedback sensor in the multi-source sensing data to obtain the target pose information.

[0072] In sub-step 23, if the confidence level of the initial pose information is lower than the preset working threshold, it indicates that the current positioning result is unreliable, for example, due to excessive inertial navigation cumulative error or insufficient environmental features leading to positioning drift.

[0073] Then, the autonomous positioning system will automatically trigger the absolute pose correction process. It uses the force data collected by the six-dimensional force feedback sensor at the end of the robotic arm and the prior information of the spatial reference object to perform absolute correction on the initial pose information, and uses the corrected initial pose information as the target pose information.

[0074] This correction can reduce the cumulative error of the inertial navigation unit by more than 90% in a single operation.

[0075] In some embodiments of this application, the contact force data is collected by the force feedback sensor when the robotic arm contacts a spatial reference object in the underwater environment. The step of correcting the initial pose information based on the contact force data collected by the force feedback sensor from the multi-source sensing data to obtain the target pose information includes: Sub-step 31: Obtain the prior information of the spatial reference object.

[0076] In sub-step 31, when the underwater robot performs a contact operation, the robotic arm comes into contact with a spatial reference object in the underwater environment. The autonomous positioning system acquires prior information about this spatial reference object, which includes at least the object's geometric dimensions, shape features, and spatial coordinate reference. This prior information can be pre-stored in the autonomous positioning system as a correction reference for subsequent absolute pose correction to obtain the target pose information.

[0077] In one example, contact-based operations could include offshore wind farm pile foundation inspection and submarine pipeline maintenance, while spatial reference objects could be standardized work objects with known geometric dimensions, such as wind turbine pile foundations and pipelines.

[0078] Sub-step 32: Based on the contact force data and the prior information, the initial pose information is corrected to obtain the target pose information.

[0079] In sub-step 32, the autonomous positioning system acquires contact force vectors and contact position information, i.e., contact force data, through a six-dimensional force feedback sensor at the end of the robotic arm. Based on this contact force data and prior information about a spatial reference object, using a standard working object of known size as an absolute spatial reference, the robot's precise pose relative to the reference object is calculated. This precise pose is then used as a correction value to absolutely correct the initial pose information obtained from the fused positioning solution, overwriting the original drift pose and obtaining the target pose information. A single correction can reduce the cumulative error of the inertial navigation unit by more than 90%, and the positioning error drift does not exceed 3cm after 72 hours of continuous operation, fundamentally solving the problem of cumulative error divergence in long-term inertial navigation operation.

[0080] In the game embodiment of this application, it also includes: Sub-step 41: Perform fault detection on the multi-source sensor.

[0081] In sub-step 41, the autonomous positioning system can monitor the operational status of the vision sensor, sonar sensor, inertial navigation unit, and force feedback sensor of the robotic arm in real time through the fault detection module in the intelligent control module. The fault detection module continuously judges whether there are faults such as data anomalies, signal loss, and communication interruptions in each sensor, ensuring that the system can detect sensor-level problems in a timely manner.

[0082] Sub-step 42: If any sensor malfunction is detected, stop using the malfunctioning sensor to collect data, and redetermine the weight allocation strategy based on the non-malfunctioning sensors.

[0083] In sub-step 42, when any sensor malfunction is detected, it is considered a Level 1 fault. The autonomous positioning system immediately stops using the data collected by the faulty sensor to prevent erroneous data from contaminating the fused positioning results. Simultaneously, the autonomous positioning system re-determines the weight allocation strategy based on the remaining normally functioning sensors. For example, when a vision sensor malfunctions, the system automatically discards visual data, increases the weight of the sonar sensor to 65%, the inertial navigation unit to 35%, and simultaneously activates the sonar feature enhancement algorithm to ensure that the positioning accuracy attenuation does not exceed 20%. This fault-tolerant mechanism, linked to the fusion weights, overcomes the limitations of simple master / slave switching in existing technologies, achieving a smooth transition of positioning capability after sensor failure.

[0084] In some embodiments of this application, it also includes: The confidence level of the target pose information is determined, and a relocation process is triggered if the confidence level of the target pose information is lower than the preset working threshold.

[0085] Specifically, after the underwater robot completes an absolute pose correction to obtain the target pose information, the autonomous positioning system evaluates the confidence level of the target pose information and determines its confidence level.

[0086] If the confidence level of the target pose information is greater than or equal to the preset working threshold, it indicates that the corrected positioning result is reliable and the system is operating normally. If the confidence level of the target pose information is still lower than the preset working threshold, it indicates that a single absolute pose correction failed to correct the initial pose information to the required operational level. At this time, the autonomous positioning system triggers a level-two fault handling mechanism, namely the autonomous repositioning process.

[0087] As an example, the triggering conditions for a Level 2 fault include two categories: one is triggered when the accuracy fails to recover after processing a Level 1 fault; the other is triggered independently by non-sensor fault causes such as exceeding positioning accuracy limits or insufficient confidence, drastic environmental changes, or excessively long calibration cycles. Moreover, after the repositioning process is triggered, the autonomous system automatically executes subsequent repositioning operations without manual intervention.

[0088] In the game embodiment of this application, the relocation process includes: Sub-step 51: Obtain the pre-stored prior map of the job scenario.

[0089] In sub-step 51, after the relocation process is triggered, the autonomous positioning system can acquire a pre-stored prior map of the operational scenario. This prior map is a pre-constructed global environmental reference map. For example, a 3D point cloud map of the offshore wind farm pile foundation area or a route map of submarine pipelines, providing an absolute spatial reference for subsequent sonar data matching.

[0090] Sub-step 52: Control the sonar sensor to acquire the relocation sonar map of the underwater environment, and match the relocation sonar map with the prior map of the operation scene to obtain the matching result.

[0091] In sub-step 52, the autonomous positioning system controls the sonar sensor to perform a panoramic scan and acquire a repositioning sonar map of the underwater environment, i.e., a local sonar image of the current environment. Then, the repositioning sonar map can be matched with a pre-stored prior map of the operation scenario. The matching algorithm finds the corresponding position of the local image in the global map. In addition, during the matching process, the matching results can be verified by combining the standard operation object feature library to ensure the accuracy of the matching and finally obtain the matching result.

[0092] The standard operation object feature library is also pre-stored in the autonomous positioning system, such as prior geometric features like pile diameter and flange spacing.

[0093] Sub-step 53: Determine the current pose information based on the matching result, and update the target pose information using the current pose information.

[0094] In sub-step 53, the autonomous localization system can calculate the underwater robot's current pose information relative to the prior map based on the matching results. The current pose information includes position and orientation. Then, the current pose information is used as the corrected absolute pose to replace the original target pose information, completing pose reconvergence. After relocalization, if the target pose information is greater than or equal to a preset working threshold, the system can exit the fault mode and continue operation without manual intervention.

[0095] In some embodiments of this application, the autonomous positioning system also includes a dynamic path planning algorithm for positioning anomalies. That is, when the target pose information is continuously lower than a preset working threshold, the autonomous positioning system can adjust its operating trajectory, prioritize moving to an underwater environment with rich features and a stable environment to complete pose recalibration, while avoiding collision risks.

[0096] It should be added that Level 1 faults can be caused by abnormal or malfunctioning single sensor data, Level 2 faults can be caused by positioning accuracy exceeding limits or insufficient confidence, and the autonomous positioning system can also include a handling mechanism for Level 3 faults. Specifically, Level 3 faults can be caused by interruption of the main communication link or abnormal power. For Level 3 faults, when the autonomous positioning system detects an interruption of the main communication link or abnormal power of the underwater robot, the autonomous positioning system can automatically switch to emergency mode, activate the backup battery and wireless acoustic backup link, shut down at least some non-essential sensors and computing load, and ensure that the core positioning function continues to operate for more than 4 hours.

[0097] As can be seen, the autonomous positioning system in this application can include a three-level fault detection mechanism, and all fault events can be simultaneously reported to the surface operator via the umbilical cable. Furthermore, during the underwater robot's operation, a comprehensive log containing environmental parameters, sensor status, weight allocation strategies, positioning accuracy, and calibration records can be generated simultaneously, supporting post-operation algorithm optimization and operational review.

[0098] In this embodiment, in response to an underwater working command, at least one environmental parameter of the underwater environment is collected at a preset sensing frequency. Multi-source sensor data is collected during the underwater working process. This multi-source sensor data is collected by multi-source sensors installed on the underwater robot, including at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor from the robotic arm. Based on the first collected environmental parameter, a weight allocation strategy is determined for the vision sensor, the sonar sensor, and the inertial navigation unit. The weight allocation strategy and the multi-source sensor data are then used to determine the optimal data allocation strategy. According to the method, the initial pose information is determined, and the target pose information is determined based on the force feedback sensor collected by the force feedback sensor in the multi-source sensing data and the initial pose information. This realizes environmental adaptive multi-sensor fusion positioning. On this basis, the initial pose information is corrected based on the force feedback sensor collected by the force feedback sensor. The contact force vector of the robotic arm is used as the absolute correction source, which effectively suppresses the cumulative error divergence of the inertial navigation unit. It can keep the positioning error drift of the underwater robot within 3cm for 72 hours of continuous operation, which is more than 50% more accurate than the positioning error of more than 10cm in the existing solution.

[0099] Furthermore, it achieves dynamic determination of weight allocation strategy through environmental parameters, breaking through the environmental limitations of single sensor schemes and fixed weight modes. It can significantly improve positioning stability under harsh conditions such as turbid waters and strong current areas, avoid data conflicts and positioning jump problems, reduce the need for human intervention, and enhance the autonomous operation capability of underwater robots.

[0100] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of this application.

[0101] Reference Figure 2 This document illustrates a structural block diagram of an autonomous positioning system for an underwater robot, provided in some embodiments of this application. The system is applied to an underwater robot and may specifically include the following modules: The environmental sensing module 201 is used to respond to underwater working instructions and collect at least one environmental parameter of the underwater environment according to a preset sensing frequency. The multi-source sensor module 202 is used to collect multi-source sensing data during underwater operation; wherein, the multi-source sensing data is collected by multi-source sensors installed on the underwater robot, and the multi-source sensors include at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor of the robotic arm. The fusion positioning module 203 is used to determine the weight allocation strategy of the visual sensor, the sonar sensor and the inertial navigation unit based on at least one environmental parameter acquired for the first time, and to determine the initial pose information based on the weight allocation strategy and the multi-source sensor data. The correction module 204 is used to determine the target pose information based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information.

[0102] The autonomous positioning system for an underwater robot provided in this application embodiment can achieve... Figure 1 The various processes of an underwater robot's autonomous localization method implemented in the method embodiments are not described in detail here to avoid repetition.

[0103] In this embodiment, in response to an underwater working command, at least one environmental parameter of the underwater environment is collected at a preset sensing frequency. Multi-source sensor data is collected during the underwater working process. This multi-source sensor data is collected by multi-source sensors installed on the underwater robot, including at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor from the robotic arm. Based on the first collected environmental parameter, a weight allocation strategy is determined for the vision sensor, the sonar sensor, and the inertial navigation unit. The weight allocation strategy and the multi-source sensor data are then used to determine the optimal data allocation strategy. According to the method, the initial pose information is determined, and the target pose information is determined based on the force feedback sensor collected by the force feedback sensor in the multi-source sensing data and the initial pose information. This realizes environmental adaptive multi-sensor fusion positioning. On this basis, the initial pose information is corrected based on the force feedback sensor collected by the force feedback sensor. The contact force vector of the robotic arm is used as the absolute correction source, which effectively suppresses the cumulative error divergence of the inertial navigation unit. It can keep the positioning error drift of the underwater robot within 3cm for 72 hours of continuous operation, which is more than 50% more accurate than the positioning error of more than 10cm in the existing solution.

[0104] Furthermore, it achieves dynamic determination of weight allocation strategy through environmental parameters, breaking through the environmental limitations of single sensor schemes and fixed weight modes. It can significantly improve positioning stability under harsh conditions such as turbid waters and strong current areas, avoid data conflicts and positioning jump problems, reduce the need for human intervention, and enhance the autonomous operation capability of underwater robots.

[0105] Optionally, this application embodiment also provides an electronic device, including a processor 310, a memory 309, and a program or instructions stored in the memory 309 and executable on the processor 310. When the program or instructions are executed by the processor 310, they implement the various processes of the above-described autonomous positioning method embodiment for underwater robots and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0106] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0107] Figure 3 A schematic diagram of the hardware structure of an electronic device to implement an embodiment of this application. The electronic device 300 includes, but is not limited to, components such as: a radio frequency unit 301, a network module 302, an audio output unit 303, an input unit 304, a sensor 305, a display unit 306, a user input unit 307, an interface unit 308, a memory 309, and a processor 310. The user input unit 307 includes a touch panel 3071 and other input devices 3072; the display unit 306 includes a display panel 3061; and the input unit includes a graphics processor 3041 and a microphone 3042.

[0108] Those skilled in the art will understand that the electronic device 300 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 310 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 3 The electronic device structure shown does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here. This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described underwater robot autonomous positioning method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0109] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0110] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described autonomous positioning method embodiment for underwater robots and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0111] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0112] It should be noted that, in this document, 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 limitations, 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 that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0113] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0114] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. An autonomous positioning method for an underwater robot, characterized in that, Applied to underwater robots, the method includes: In response to underwater work instructions, at least one environmental parameter of the underwater environment is collected according to a preset sensing frequency; Collect multi-source sensor data during underwater operation; wherein, the multi-source sensor data is collected by multi-source sensors installed on the underwater robot, and the multi-source sensors include at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor of the robotic arm; Based on at least one environmental parameter acquired initially, a weight allocation strategy for the visual sensor, the sonar sensor, and the inertial navigation unit is determined, and initial pose information is determined based on the weight allocation strategy and the multi-source sensor data. Based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information, the target pose information is determined.

2. The method according to claim 1, characterized in that, The multi-source sensing data includes initial pixel images and initial sonar images acquired by the visual sensor and the sonar sensor at a first frequency, and inertial navigation data acquired by the inertial navigation unit at a second frequency. The step of determining initial pose information based on the weight allocation strategy and the multi-source sensing data includes: Based on a preset filtering mechanism, feature point filtering processing is performed on the initial pixel image and the initial sonar image to obtain the target pixel image and the target sonar image. According to the weight allocation strategy, the target pixel map and the target sonar map at the first frequency and the inertial navigation data at the second frequency are fused together, and the initial pose information is output according to the preset output frequency.

3. The method according to claim 1, characterized in that, Based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information, the target pose information is determined, including: Determine the confidence level of the initial pose information; If the confidence level of the initial pose information is greater than or equal to a preset working threshold, the initial pose information is used as the target pose information. If the confidence level of the initial pose information is lower than the preset working threshold, the initial pose information is corrected based on the tactile force data collected by the force feedback sensor in the multi-source sensing data to obtain the target pose information.

4. The method according to claim 3, characterized in that, The contact force data is collected by the force feedback sensor when the robotic arm comes into contact with a spatial reference object in the underwater environment. The step of correcting the initial pose information based on the contact force data collected by the force feedback sensor to obtain the target pose information includes: Obtain prior information about the spatial reference object; Based on the contact force data and the prior information, the initial pose information is corrected to obtain the target pose information.

5. The method according to any one of claims 1-4, characterized in that, Also includes: The weight allocation strategy is adjusted based on at least one currently collected environmental parameter and a preset adjustment range.

6. The method according to any one of claims 1-4, characterized in that, Also includes: Fault detection is performed on the multi-source sensors; If any sensor malfunction is detected, data collection using the malfunctioning sensor is stopped, and the weight allocation strategy is redefined based on the non-malfunctioning sensors.

7. The method according to claim 3, characterized in that, Also includes: The confidence level of the target pose information is determined, and a relocation process is triggered if the confidence level of the target pose information is lower than the preset working threshold.

8. The method according to claim 7, characterized in that, The relocation process includes: Obtain the pre-stored prior map of the job scenario; The sonar sensor is controlled to acquire a relocation sonar map of the underwater environment, and the relocation sonar map is matched with a prior map of the operation scenario to obtain a matching result; Based on the matching result, the current pose information is determined, and the target pose information is updated using the current pose information.

9. The method according to any one of claims 1-4, characterized in that, The underwater robot includes multiple preset working modes for different scenarios. These preset working modes include an inspection mode, a precision operation mode, and an emergency mode. The inspection mode, the precision operation mode, and the emergency mode each have corresponding weight allocation strategies and preset output frequencies.

10. An autonomous positioning system for an underwater robot, characterized in that, The system, applied to underwater robots, includes: The environmental sensing module is used to respond to underwater working instructions and collect at least one environmental parameter of the underwater environment according to a preset sensing frequency. A multi-source sensor module is used to collect multi-source sensing data during underwater operation; wherein, the multi-source sensing data is collected by multi-source sensors installed on the underwater robot, and the multi-source sensors include at least a vision sensor, a sonar sensor, an inertial navigation unit, and a force feedback sensor of the robotic arm; The fusion positioning module is used to determine the weight allocation strategy of the visual sensor, the sonar sensor and the inertial navigation unit based on at least one environmental parameter acquired for the first time, and to determine the initial pose information based on the weight allocation strategy and the multi-source sensor data. The correction module is used to determine the target pose information based on the tactile force data collected by the force feedback sensor in the multi-source sensing data and the initial pose information.

11. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the method as described in any one of claims 1-9.

12. A readable storage medium, characterized in that, A computer program is stored on the readable storage medium, which, when executed by a processor, implements the method as described in any one of claims 1-9.

Images