Control method and system for preventing umbilical cord winding of underwater cleaning robot

Abstract

According to the method, data are comprehensively collected by means of a multi-source sensor, training is carried out by adopting a deep learning model fusing a long-short-term memory network (LSTM) and a convolutional neural network (CNN) after rigorous links such as data preprocessing and feature extraction, and the umbilical cord winding of the underwater cleaning robot is prevented. The system is composed of a sensor module, a data processing module, a machine learning module and a control execution module in a cooperative mode, by means of the achievement of the invention, the real-time state of the umbilical cable can be sensitively and accurately sensed, then the motion trail of the underwater cleaning robot is intelligently regulated and controlled, and the underwater cleaning robot can be accurately positioned and controlled. According to the underwater cleaning device, the safety of underwater cleaning operation is effectively enhanced, the operation efficiency is greatly improved, and a technical breakthrough with milestone significance is brought to the field of underwater cleaning.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater robot control, and in particular to a control method and system for preventing the umbilical cord of an underwater cleaning robot from being entangled. Background Art

[0002] With the increasing demand for marine resource development and underwater facility maintenance, the application of underwater cleaning robots is becoming increasingly widespread. Underwater cleaning robots are typically connected to surface support systems via umbilical cables for power and communication. However, in complex underwater environments, umbilical cables are susceptible to factors such as currents and the adhesion of marine organisms, leading to entanglement and pulling, which can restrict the robot's movement and even cause safety accidents.

[0003] Existing umbilical cable positioning control technologies often rely on fixed, pre-set programs or simple feedback mechanisms, making them difficult to quickly adapt to the dynamic changes in complex underwater environments. For example, in turbulent water, traditional control methods cannot promptly adjust the umbilical cable tension and robot motion, which can easily lead to umbilical cable breakage or robot loss of control, seriously affecting operational efficiency and equipment safety. Therefore, a more intelligent and efficient umbilical cable positioning control solution is urgently needed. Summary of the Invention

[0004] The present invention aims to provide a control method and system to prevent the umbilical cord of an underwater cleaning robot from being entangled, thereby realizing accurate perception and intelligent regulation of the umbilical cable status, and improving the stability, safety and efficiency of the underwater cleaning robot operating in complex environments.

[0005] In order to solve the above problems, the present invention provides a technical solution:

[0006] A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot, comprising the following steps:

[0007] S1. Utilize pressure sensors, tension sensors, gyroscopes, accelerometers, and high-definition cameras deployed at key locations on the underwater cleaning robot and umbilical cable to collect multi-source data related to the umbilical cable status and underwater environment in real time and transmit it to the data processing module.

[0008] S2, the data processing module uses the Kalman filter algorithm to reduce noise of numerical data such as pressure and tension, uses median filtering and histogram equalization to process image data, and normalizes all data;

[0009] S3. Use machine learning algorithms to extract features from the preprocessed data. The time and frequency domain features of the pressure and tension data, the dynamic features of the posture data, and the visual features extracted from the image data using an improved CNN combined with an attention mechanism are integrated to form a multidimensional feature vector.

[0010] S4. Build a deep learning hybrid model that integrates LSTM and CNN, use historical data covering different underwater environments and operation scenarios for training, use feature vectors as input, and use the optimal control strategy as output to optimize model parameters;

[0011] S5. The feature vectors collected and processed in real time are input into the trained model. The model outputs the real-time position, status assessment results and control instructions of the umbilical cable. The control execution module adjusts the robot movement and the retraction and extension of the umbilical cable according to the instructions.

[0012] Preferably, the pressure sensor is used to monitor the water pressure at different parts of the umbilical cable, and the tension sensor is used to obtain real-time tension data of each section of the umbilical cable.

[0013] Preferably, the dynamic characteristics of the posture data include the robot's rotational angular velocity and acceleration direction change rate.

[0014] Preferably, the visual features of the image data include the shape, position, and relative distance from obstacles of the umbilical cable.

[0015] Preferably, the historical data includes sensor data under normal operation, umbilical cable entanglement, and water flow impact conditions and corresponding expert control strategies.

[0016] An intelligent positioning and control system for an umbilical cable of an underwater cleaning robot based on machine learning includes: a sensor module composed of a pressure sensor, a tension sensor, a gyroscope, an accelerometer, and a high-definition camera, for collecting data related to the umbilical cable and the underwater environment and transmitting it to a data processing module;

[0017] Data processing module, used for noise reduction, filtering, normalization and feature extraction of sensor data;

[0018] The machine learning module is used to build and train a deep learning hybrid model that integrates LSTM and CNN, receives feature vectors, and outputs umbilical cable positioning information and control instructions;

[0019] The control execution module is used to control the robot motion mechanism and the umbilical cable retracting and releasing device according to the control instructions.

[0020] Preferably, the data processing module uses a Kalman filter algorithm to process numerical data, and uses median filtering and histogram equalization to process image data.

[0021] Preferably, the machine learning module uses historical data covering different underwater environments and operation scenarios to train the model.

[0022] Preferably, the control execution module controls the robot movement by adjusting the robot propeller speed and the steering angle of the servo, and controls the umbilical cable tension and length by adjusting the umbilical cable receivable and discharge motor speed and steering.

[0023] Preferably, the data processing module, machine learning module and control execution module are integrated on a dedicated control board, and the data processing module and the control board exchange data via a high-speed SPI bus.

[0024] The beneficial effects of the present invention are:

[0025] Precise positioning and real-time monitoring: The combination of multi-source sensors and advanced feature extraction algorithms can obtain the umbilical cable position and status information with high precision, realizing real-time dynamic monitoring.

[0026] Intelligent decision-making and adaptive control: The deep learning hybrid model is trained with large amounts of data to quickly analyze complex underwater environments, intelligently generate control strategies, and effectively respond to situations such as water flow changes and obstacle interference.

[0027] Improve operational safety and reliability: Timely and accurate control avoids umbilical cable failures, reduces equipment damage risks, and ensures safe and stable underwater operations.

[0028] Improved operational efficiency: Optimized control strategies reduce robot downtime caused by umbilical cable issues and improve cleaning efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] For ease of explanation, the present invention is described in detail with reference to the following specific implementations and accompanying drawings.

[0030] Figure 1 It is a schematic diagram of the workflow of the present invention. DETAILED DESCRIPTION

[0031] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present disclosure. Rather, they are merely examples of devices consistent with certain aspects of the present disclosure, as detailed in the appended claims.

[0032] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0033] like Figure 1 As shown, this specific embodiment adopts the following technical solutions:

[0034] Example:

[0035] A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot, comprising the following steps:

[0036] Data acquisition: Pressure sensors, tension sensors, gyroscopes, accelerometers, and high-definition cameras are deployed at key locations on the underwater cleaning robot and the umbilical cable. The pressure sensors monitor the water pressure at different locations on the umbilical cable in real time. The tension sensors obtain real-time tension data on each section of the umbilical cable. The gyroscopes and accelerometers detect the robot's posture and motion status. The high-definition camera captures images of the underwater environment around the umbilical cable. Multi-source data is synchronously transmitted to the data processing module.

[0037] Data preprocessing: The data processing module performs noise reduction, filtering, and normalization on the collected data. For numerical data such as pressure and tension, the Kalman filter algorithm is used to remove noise. For image data, median filtering and histogram equalization are used to enhance image clarity and normalize all data to a unified numerical range.

[0038] Feature extraction: Machine learning algorithms are used to extract features from preprocessed data. For pressure and tension data, time domain features (such as mean, variance, and kurtosis) and frequency domain features (spectral features are obtained through Fourier transform) are extracted. For posture data, dynamic features such as the robot's rotational angular velocity and rate of change of acceleration are extracted. For image data, an improved convolutional neural network (CNN) structure is used in conjunction with an attention mechanism to extract visual features such as the umbilical cable's shape, position, and relative distance to obstacles. These features are then fused into a multidimensional feature vector.

[0039] Model training: A hybrid model based on deep learning is constructed, integrating a long short-term memory network (LSTM) and a convolutional neural network (CNN). This model is trained using a large amount of historical data covering various underwater environments and operating scenarios. The data includes sensor data under various operating conditions, such as normal operation, umbilical cable entanglement, and water flow impact, as well as corresponding expert control strategies. The fused feature vector is used as input, and the optimal control strategy is used as output. The model parameters are optimized through a back-propagation algorithm, enabling the model to predict the optimal control strategy based on the input data.

[0040] Real-time positioning and control: The feature vectors collected and processed in real time are input into the trained model. The model outputs the real-time position, status evaluation results and corresponding control instructions of the umbilical cable. The control execution module coordinates and adjusts the propeller and steering gear of the underwater cleaning robot according to the instructions, controls the movement posture and speed of the robot, and drives the umbilical cable retraction device to adjust the umbilical cable tension and length to achieve intelligent positioning control.

[0041] Based on the above method, the present invention also provides a system for intelligent positioning control of the umbilical cable of an underwater cleaning robot based on machine learning, the system comprising:

[0042] Sensor module: It consists of pressure sensor, tension sensor, gyroscope, accelerometer and high-definition camera, responsible for collecting data related to the umbilical cable and underwater environment in real time and transmitting it to the data processing module.

[0043] Data processing module: pre-processes and extracts features from sensor data, converting raw data into feature vectors suitable for machine learning model processing;

[0044] Machine Learning Module: Builds and trains a deep learning hybrid model, receives feature vectors, and outputs umbilical cable positioning information and control instructions;

[0045] Control execution module: According to the control instructions, it controls the motion mechanism of the underwater cleaning robot and the umbilical cable retraction and release device to realize intelligent positioning control of the umbilical cable.

[0046] Specifically:

[0047] High-precision pressure sensors and tension sensors are installed at appropriate locations on the underwater cleaning robot's shell to ensure they can accurately sense the stress on the umbilical cable. Gyroscopes and accelerometers are fixed to the robot's center of gravity to ensure accurate posture detection. A high-definition camera is installed at the front of the robot, covering the umbilical cable connection area and the forward working environment. Each sensor is connected to the data processing module via a waterproof data cable. The data processing module uses a high-performance ARM processor with powerful data processing and computing capabilities. The machine learning module and control execution module are integrated on a dedicated control board, and the control board and data processing module exchange data via a high-speed SPI bus.

[0048] In the software design of the data processing module, C++ is used to write a data preprocessing program, implement the Kalman filter algorithm to reduce noise for numerical data, and use OpenCV library functions to complete median filtering and histogram equalization of image data. For feature extraction, Python is combined with the Scikit-learn library to extract numerical data features. The TensorFlow framework is used to build an improved convolutional neural network to extract image features, and a program is written to implement feature fusion. The machine learning module builds a deep learning model that integrates LSTM and CNN based on Python and TensorFlow frameworks. By collecting a large amount of data generated by actual underwater operations and simulation experiments, the model is divided into training sets, validation sets, and test sets. The model is trained and optimized, and appropriate hyperparameters such as learning rate and batch size are set. After multiple rounds of iterative training, the model achieves optimal performance. The software of the control execution module writes the corresponding driver according to the control instructions output by the machine learning module. The robot motion control is achieved by controlling the speed of the robot's thrusters and the steering angle of the servo. The tension and length of the umbilical cable are controlled by adjusting the speed and direction of the umbilical cable's retractable motor.

[0049] When the underwater cleaning robot performs the task of cleaning cages in offshore aquaculture areas, the sensor module collects real-time pressure and tension data of the umbilical cable under the action of water flow, as well as the posture data and surrounding environment images of the robot when avoiding aquaculture facilities. After the data processing module preprocesses the data and extracts features, it inputs the feature vector into the machine learning module. If the model detects that the umbilical cable is at risk of being entangled in the aquaculture cage, it outputs a control instruction. The control execution module then controls the robot to change its direction of movement and appropriately tightens the umbilical cable to avoid entanglement, ensuring that the robot successfully completes the cage cleaning task.

[0050] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

[0051] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope of the present disclosure is indicated by the following claims.

Claims

1. A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot, characterized in that: The following steps are involved: S1. Utilize pressure sensors, tension sensors, gyroscopes, accelerometers, and high-definition cameras deployed at key locations on the underwater cleaning robot and umbilical cable to collect multi-source data related to the umbilical cable status and underwater environment in real time and transmit it to the data processing module. S2, the data processing module uses the Kalman filter algorithm to reduce noise of numerical data such as pressure and tension, uses median filtering and histogram equalization to process image data, and normalizes all data; S3. Use machine learning algorithms to extract features from the preprocessed data. The time and frequency domain features of the pressure and tension data, the dynamic features of the posture data, and the visual features extracted from the image data using an improved CNN combined with an attention mechanism are integrated to form a multidimensional feature vector. S4. Build a deep learning hybrid model that integrates LSTM and CNN, use historical data covering different underwater environments and operation scenarios for training, use feature vectors as input, and use the optimal control strategy as output to optimize model parameters; S5. The feature vectors collected and processed in real time are input into the trained model. The model outputs the real-time position, status assessment results and control instructions of the umbilical cable. The control execution module adjusts the robot movement and the retraction and extension of the umbilical cable according to the instructions.

2. A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 1, characterized in that: The pressure sensor is used to monitor the water pressure on different parts of the umbilical cable, and the tension sensor is used to obtain real-time tension data of each section of the umbilical cable.

3. A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 1, characterized in that: The dynamic characteristics of the posture data include the robot's rotational angular velocity and acceleration direction change rate.

4. A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 1, characterized in that: The visual features of the image data include the shape, position, and relative distance from obstacles of the umbilical cable.

5. A control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 1, characterized in that: The historical data includes sensor data under normal operation, umbilical cable entanglement, and water flow impact conditions and corresponding expert control strategies.

6. An intelligent positioning control system for the umbilical cable of an underwater cleaning robot based on machine learning, characterized in that: include: The sensor module, consisting of a pressure sensor, a tension sensor, a gyroscope, an accelerometer, and a high-definition camera, is used to collect data related to the umbilical cable and the underwater environment and transmit it to the data processing module; Data processing module, used for noise reduction, filtering, normalization and feature extraction of sensor data; The machine learning module is used to build and train a deep learning hybrid model that integrates LSTM and CNN, receives feature vectors, and outputs umbilical cable positioning information and control instructions; The control execution module is used to control the robot motion mechanism and the umbilical cable retracting and releasing device according to the control instructions.

7. The control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 6 is characterized in that: The data processing module uses the Kalman filter algorithm to process numerical data and uses median filtering and histogram equalization to process image data.

8. The control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 6 is characterized in that: The machine learning module uses historical data covering different underwater environments and operation scenarios to train the model.

9. The control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 6 is characterized in that: The control execution module controls the robot movement by adjusting the robot propeller speed and the steering angle of the steering gear, and controls the umbilical cable tension and length by adjusting the umbilical cable receivable and discharge motor speed and steering.

10. The control method and system for preventing umbilical cord entanglement of an underwater cleaning robot according to claim 6 is characterized in that: The data processing module, machine learning module and control execution module are integrated on a dedicated control board, and the data processing module and the control board exchange data via a high-speed SPI bus.

Images