ROV motion control system and method based on PID

Abstract

The invention relates to the technical field of control systems, provides an ROV motion control system and method based on PID, and solves the defects of an existing PID control algorithm in ROV complex system control, the system comprises a data acquisition module, a data processing module, a data detection module, a model matching module and an optimization module, historical data are acquired and an algorithm model knowledge base is constructed, and the ROV motion control algorithm is optimized. The method comprises the steps that environment data are matched in real time, an optimal control model is selected, a control command is sent, ROV operation data are collected, if real-time data are not consistent with expected data, the control model is optimized through a genetic algorithm, and an optimized control command is generated. And the robustness and the operation efficiency of the system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of control systems, and in particular to a ROV motion control system and method based on PID. Background Art

[0002] PID, which stands for Proportional, Integral, and Derivative control, is a control algorithm. The PID control algorithm has the advantages of simple principle, easy implementation, wide application, independent control parameters, and relatively simple parameter selection. The core of the PID controller is to calculate the deviation between the given value and the actual value, and linearly combine this deviation through three different adjustment functions: proportional, integral, and differential, to generate a control signal to control the controlled object. ROV, or Remotely Operated Vehicle, is a device that can move underwater, has a visual and perception system, and uses a manipulator or other tools through remote control or autonomous operation to replace or assist people in completing underwater operations.

[0003] The PID-based ROV motion control system is an underwater ROV robot control solution that uses the PID (proportional-integral-differential) control algorithm to calculate the deviation between the current position and the target position of the ROV (remote control underwater ROV robot) in real time, and generates control instructions through a linear combination of the three regulation functions of proportion, integration and differentiation. The control instructions are used to adjust the thruster output of the ROV, thereby controlling its posture and motion trajectory underwater.

[0004] The existing PID ROV motion control system has the following technical pain points in actual operation: ROV is affected by multiple factors such as water flow, waves, load changes, etc. when it moves underwater. The influence of multiple factors such as water flow, waves, load changes, etc. is nonlinear and time-varying. Existing PID control algorithms usually assume that the system behavior is linear and time-invariant, resulting in PID control being unable to provide accurate control effects when facing complex systems such as ROV. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a ROV motion control system and method based on PID, which solves the problem that the existing PID control algorithm usually assumes that the system behavior is linear and time-invariant, resulting in the PID control being unable to provide accurate control effects when facing a complex system such as an ROV.

[0006] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows: In a first aspect, the present invention provides a ROV motion control system based on PID, comprising: A data acquisition module is used to acquire the historical data of the motion control system, which includes the target position historical data, the actual position historical data, the deviation historical data, the control parameter historical data, the environment historical data, the sensor historical data, the motion trajectory historical data and the control instruction historical data; A data processing module is used to extract data features from environmental data to obtain environmental data classification features, match the environmental classification features with historical data of the motion control system based on time information to obtain an environmental classification feature data set, divide each type of environmental feature data in the environmental classification feature data set into a training set and a verification set, and use each type of environmental feature data divided into a training set and a verification set to train a neural network model to obtain a motion control model data set corresponding to the environmental classification features, and construct the motion control model data set corresponding to the environmental classification features into an algorithm model knowledge base; A data detection module is used to receive real-time target location data and real-time water flow data, real-time wave data and real-time load data, obtain target location natural environment data according to the real-time target location data, and aggregate the real-time water flow data, real-time wave data, real-time load data and target location natural environment data to obtain a real-time environment data set; The model matching module is used to match the real-time environment data set in the preset algorithm model knowledge base, obtain the motion control model corresponding to the real-time environment data, receive the real-time control command, substitute the real-time control command into the motion control model corresponding to the real-time environment data, and obtain the expected operation data of the ROV robot; The model optimization module is used to transmit the real-time control command to the ROV robot, collect the real-time operation data of the ROV robot, and compare the real-time operation data of the ROV robot with the expected operation data of the ROV robot. If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the motion control model corresponding to the real-time environment data is optimized using a genetic algorithm according to the error data between the real-time operation data of the ROV robot and the expected operation data of the ROV robot to obtain the motion control model corresponding to the optimized real-time environment data, and the real-time control command is processed using the motion control model corresponding to the optimized real-time environment data to generate an optimized ROV robot control command, and the optimized ROV robot control command is sent to the ROV robot.

[0007] Furthermore, in the PID-based ROV motion control system of the present invention, the data detection module is also used for: The specific steps of matching the real-time environment data set in the preset algorithm model knowledge base to obtain the motion control model corresponding to the real-time environment data include: Extract features from real-time environmental data sets to obtain water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features; The extracted features are combined into feature vectors according to the preset format and order. The feature vectors will be used for searching and matching in the algorithm model knowledge base. The feature vector is used to search in the algorithm model knowledge base to obtain the motion control model that best matches the real-time environment data. The motion control model that best matches the real-time environment data is used to calculate the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base; According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

[0008] Furthermore, in the PID-based ROV motion control system of the present invention, the model matching module is also used for: Sending real-time control commands to the ROV robot through a communication link, the real-time control commands including target position commands, speed commands and attitude commands; After the ROV robot receives the control command, the internal control system of the ROV robot interprets the command and drives the corresponding actuator to move; In the process of the ROV robot executing the control command, the operating data of the ROV robot is collected in real time through the sensors installed on the ROV robot. The sensors on the ROV robot include position sensors, speed sensors and attitude sensors. The operating data of the ROV robot include actual position, actual speed, actual attitude, motor current and voltage; Recalculate the expected operation data of the ROV robot predicted by the motion control model based on real-time environmental data and control commands; The preprocessed real-time operation data of the ROV robot is compared with the expected operation data item by item, and the comparison results are output to the model optimization module.

[0009] Furthermore, in the PID-based ROV motion control system of the present invention, the model optimization module is also used for: If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the specific steps of using a genetic algorithm to optimize the motion control model corresponding to the real-time environmental data and generate an optimized ROV robot control command include: Calculate error data, which includes position error, velocity error and attitude error; Taking reducing the error as the optimization goal, define the fitness function; Initialize the genetic algorithm and set the parameters of the genetic algorithm. Each individual represents a set of motion control model parameters. Use each individual in the population, i.e., the model parameter set, to simulate the real-time control command and obtain the simulated ROV robot operation data; Calculate the error between the simulated data and the real-time running data, and use it to evaluate the fitness of each individual; According to the fitness, excellent individuals are selected as parents, crossover operations are performed to generate offspring, and mutation operations are performed on the offspring. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirement is reached; Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data; Use the optimized motion control model to process the real-time control commands and obtain the expected operation data of the ROV robot; Based on the optimized expected operation data, an optimized ROV robot control command is generated, and the optimized ROV robot control command is sent to the ROV robot through a communication link to adjust the ROV robot motion state so that the ROV robot reaches the expected operation state.

[0010] In a second aspect, the present invention provides a PID-based ROV motion control method, which is applied to the PID-based ROV motion control system, comprising: Step S101, acquiring motion control system historical data, the motion control system historical data including target position historical data, actual position historical data, deviation historical data, control parameter historical data, environment historical data, sensor historical data, motion trajectory historical data and control instruction historical data; Step S102, extracting data features from the environmental data to obtain environmental data classification features, matching the environmental classification features with the motion control system historical data based on time information to obtain an environmental classification feature data set, dividing each type of environmental feature data in the environmental classification feature data set into a training set and a verification set, and using each type of environmental feature data divided into a training set and a verification set to train a neural network model to obtain a motion control model data set corresponding to the environmental classification features, and constructing the motion control model data set corresponding to the environmental classification features into an algorithm model knowledge base; Step S103, receiving real-time target location data and real-time water flow data, real-time wave data and real-time load data, acquiring target location natural environment data according to the real-time target location data, and aggregating the real-time water flow data, real-time wave data, real-time load data and target location natural environment data to obtain a real-time environment data set; Step S104, matching the real-time environment data set in a preset algorithm model knowledge base to obtain a motion control model corresponding to the real-time environment data, receiving a real-time control command, substituting the real-time control command into the motion control model corresponding to the real-time environment data, and obtaining the expected operation data of the ROV robot; Step S105, transmit the real-time control command to the ROV robot, collect the real-time operation data of the ROV robot, compare the real-time operation data of the ROV robot with the expected operation data of the ROV robot, if the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, then use a genetic algorithm to optimize the motion control model corresponding to the real-time environment data according to the error data between the real-time operation data of the ROV robot and the expected operation data of the ROV robot, obtain the motion control model corresponding to the optimized real-time environment data, use the motion control model corresponding to the optimized real-time environment data to perform data processing on the real-time control command, generate an optimized ROV robot control command, and send the optimized ROV robot control command to the ROV robot.

[0011] Furthermore, in the PID-based ROV motion control method of the present invention, step S103 comprises: The specific steps of matching the real-time environment data set in the preset algorithm model knowledge base to obtain the motion control model corresponding to the real-time environment data include: Extract features from real-time environmental data sets to obtain water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features; The extracted features are combined into feature vectors according to the preset format and order. The feature vectors will be used for searching and matching in the algorithm model knowledge base. The feature vector is used to search in the algorithm model knowledge base to obtain the motion control model that best matches the real-time environment data. The motion control model that best matches the real-time environment data is used to calculate the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base; According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

[0012] Furthermore, in the PID-based ROV motion control method of the present invention, step S104 comprises: Sending real-time control commands to the ROV robot through a communication link, the real-time control commands including target position commands, speed commands and attitude commands; After the ROV robot receives the control command, the internal control system of the ROV robot interprets the command and drives the corresponding actuator to move; In the process of the ROV robot executing the control command, the operating data of the ROV robot is collected in real time through the sensors installed on the ROV robot. The sensors on the ROV robot include position sensors, speed sensors and attitude sensors. The operating data of the ROV robot include actual position, actual speed, actual attitude, motor current and voltage; Recalculate the expected operation data of the ROV robot predicted by the motion control model based on real-time environmental data and control commands; The preprocessed real-time operation data of the ROV robot is compared with the expected operation data item by item, and the comparison results are output to the model optimization module.

[0013] Furthermore, in the PID-based ROV motion control method of the present invention, step S105 comprises: If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the specific steps of using a genetic algorithm to optimize the motion control model corresponding to the real-time environmental data and generate an optimized ROV robot control command include: Calculate error data, which includes position error, velocity error and attitude error; Taking reducing the error as the optimization goal, define the fitness function; Initialize the genetic algorithm and set the parameters of the genetic algorithm. Each individual represents a set of motion control model parameters. Use each individual in the population, i.e., the model parameter set, to simulate the real-time control command and obtain the simulated ROV robot operation data; Calculate the error between the simulated data and the real-time running data, and use it to evaluate the fitness of each individual; According to the fitness, excellent individuals are selected as parents, crossover operations are performed to generate offspring, and mutation operations are performed on the offspring. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirement is reached; Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data; Use the optimized motion control model to process the real-time control commands and obtain the expected operation data of the ROV robot; Based on the optimized expected operation data, an optimized ROV robot control command is generated, and the optimized ROV robot control command is sent to the ROV robot through a communication link to adjust the ROV robot motion state so that the ROV robot reaches the expected operation state.

[0014] Beneficial effects of the present invention: By dynamically selecting and optimizing the motion control model, the present invention can more accurately predict and control the motion state of the ROV robot, reduce position error, speed error and attitude error, and thus improve control accuracy. The present invention uses historical data and real-time environmental data to build an algorithm model knowledge base, selects the most appropriate motion control model for different environmental conditions, such as water flow velocity, wave height, load weight, etc., and enhances the system's adaptability to complex and changeable underwater environments.

[0015] By optimizing the motion control model through genetic algorithms, the present invention can continuously learn and adjust model parameters to adapt to new environments and control requirements, thereby improving the robustness of the system. The present invention continuously compares the real-time operation data and expected operation data of the ROV robot during real-time operation, optimizes the control model according to the error data, thereby generating more reasonable control commands and optimizing the control strategy.

[0016] Through precise control and optimized motion trajectory, the present invention can enable the ROV robot to reach the target position more quickly, complete underwater operation tasks, and improve operation efficiency. By optimizing control commands and reducing unnecessary movements, the present invention can reduce the energy consumption of the ROV robot and extend its underwater operation time.

[0017] The present invention combines multiple advanced technologies such as PID control algorithm, neural network model, genetic algorithm, etc., provides new ideas and methods for the intelligent control of ROV robots, and promotes the intelligent development of underwater robot technology.

[0018] In summary, the present invention significantly improves the control accuracy, environmental adaptability and system robustness of the ROV robot in a complex underwater environment by dynamically selecting and optimizing the motion control model, while optimizing the control strategy, improving the operating efficiency and reducing the energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on the drawings without paying any creative labor.

[0020] Figure 1 A flow chart of a PID-based ROV motion control method provided in an embodiment of the present invention.

[0021] Figure 2 A module schematic diagram of a PID-based ROV motion control system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. The technical solutions provided by the embodiments of the present invention are described in detail below in conjunction with the drawings. In order to better understand the purpose of the present invention, the present invention is further described in detail below.

[0023] First, see Figure 2 The present invention provides a ROV motion control system based on PID, comprising: A data acquisition module acquires historical data of the motion control system, which includes target position historical data, actual position historical data, deviation historical data, control parameter historical data, environment historical data, sensor historical data, motion trajectory historical data, and control instruction historical data; Target position history data: records the target position information that the ROV robot was instructed to reach in the past.

[0024] Actual position history data: records the actual position information reached by the ROV robot during actual operation.

[0025] Deviation history data: Calculates and records the difference between the target position and the actual position, reflecting the accuracy of the control system.

[0026] Control parameter history data: records the various parameter values ​​used to control the ROV robot, such as the proportional, integral, and differential coefficients of the PID controller.

[0027] Environmental history data: records various environmental information encountered by the ROV robot during its past operations, such as water flow speed, wave height, load weight, water temperature, salinity, and seabed topography.

[0028] Sensor historical data: collects historical data from various sensors on the ROV robot, such as position sensors, speed sensors, attitude sensors, etc. These data provide a detailed record of the operating status of the ROV robot.

[0029] Motion trajectory history data: records the complete motion trajectory of the ROV robot during its past operation, which helps to analyze the ROV's motion mode and performance.

[0030] Control command history data: records all control commands sent to the ROV robot, including target position commands, speed commands, attitude commands, etc.

[0031] Support model training: The collected historical data provides rich samples for subsequent neural network model training, which helps the model learn the control laws in various environments.

[0032] Improve control accuracy: By comparing historical data of target position and actual position, the accuracy of the control system can be evaluated and the control strategy can be adjusted accordingly to improve control accuracy.

[0033] Enhance system adaptability: Environmental historical data and sensor historical data provide the system with an understanding of complex environments, helping the system make more reasonable control decisions when facing new environments.

[0034] Optimize control parameters: The historical data of control parameters provides a basis for adjusting and optimizing the parameters of the PID controller, which helps to improve the performance of the control system.

[0035] The data processing module extracts data features from the environmental data to obtain environmental data classification features, matches the environmental classification features with the historical data of the motion control system according to time information to obtain an environmental classification feature data set, divides each type of environmental feature data in the environmental classification feature data set into a training set and a verification set, and uses each type of environmental feature data divided into a training set and a verification set to train a neural network model to obtain a motion control model data set corresponding to the environmental classification features, and constructs the motion control model data set corresponding to the environmental classification features into an algorithm model knowledge base; The data processing module first extracts features from historical environmental data, which includes factors such as water flow velocity, wave height, load weight, water temperature, salinity, and seabed topography.

[0036] Through feature extraction technology, the data processing module identifies key feature information from the original environmental data, and the feature information will be used for subsequent data classification and model training. Based on the extracted environmental data features, the data processing module further classifies the data to form environmental data classification features. The purpose of classification is to match the historical data of the motion control system with the corresponding environmental conditions, so as to more accurately reflect the motion control characteristics of the ROV robot in different environments.

[0037] The data processing module matches the environmental classification features with the historical data of the motion control system based on time information. Through matching, the data processing module can construct multiple environmental classification feature data sets, each of which contains the operating data and control parameters of the ROV robot under specific environmental conditions.

[0038] For each environmental classification feature dataset, the data processing module divides it into a training set and a validation set. The training set is used to train the neural network model, while the validation set is used to evaluate the performance of the model.

[0039] Using the divided data sets, the data processing module trains the neural network model. Through continuous learning and optimization, the model can learn how to effectively control the movement of the ROV robot under different environmental conditions.

[0040] After training, the data processing module collects the motion control model data corresponding to each environmental classification feature and builds an algorithm model knowledge base. The algorithm model knowledge base contains motion control models under various environmental conditions, and can quickly select the most appropriate control model according to the current environmental conditions during real-time control.

[0041] A data detection module receives real-time target location data and real-time water flow data, real-time wave data and real-time load data, obtains target location natural environment data according to the real-time target location data, and aggregates the real-time water flow data, real-time wave data, real-time load data and target location natural environment data to obtain a real-time environment data set; The data detection module receives data from different sources in real time, including real-time target position data, real-time water flow data, real-time wave data, and real-time load data. The data is transmitted to the data detection module in real time through sensors, communication systems or other data interfaces.

[0042] Based on the received real-time target location data, the data detection module will further acquire the natural environment data of the location.

[0043] Natural environmental data involves querying pre-stored environmental databases, or real-time requests for environmental information from other systems (such as marine environmental monitoring systems). Natural environmental data include water temperature, salinity, seabed topography, water flow speed, wave height, etc.

[0044] The data detection module summarizes and integrates the received real-time water flow data, real-time wave data, real-time load data, and the acquired natural environment data of the target location. Through data cleaning, format conversion, and unit unification, the accuracy of all data is ensured. The integrated data forms a real-time environmental data set, which includes all key environmental factors that the ROV robot will encounter currently and in the future. The real-time environmental data set is passed to other modules of the system (such as the model matching module) for further processing and analysis.

[0045] The model matching module matches the real-time environment data set in the preset algorithm model knowledge base to obtain the motion control model corresponding to the real-time environment data, receives the real-time control command, substitutes the real-time control command into the motion control model corresponding to the real-time environment data, and obtains the expected operation data of the ROV robot; The model matching module receives the real-time environment data set from the data detection module. In the preset algorithm model knowledge base, the module matches according to the characteristics of the real-time environment data set to find the motion control model that best matches the current environmental conditions.

[0046] The algorithm model knowledge base usually includes motion control models under a variety of different environmental conditions. The motion control model is obtained through training with historical data and can reflect the impact of different environments on the motion control of the ROV robot.

[0047] Through the matching process, the model matching module selects the motion control model that best matches the real-time environment data set from the algorithm model knowledge base. The best matching motion control model will be used for subsequent control command processing and expected operation data generation.

[0048] The model matching module receives real-time control commands from the control system, which include target position, speed, attitude and other requirements. The module substitutes the real-time control commands into the selected motion control model and generates the expected operation data of the ROV robot through calculation and reasoning of the model. The expected operation data includes the expected position, speed, attitude, thrust and other key parameters of the ROV robot.

[0049] The model matching module outputs the generated expected operation data to the actuator of the ROV robot to guide it to perform actual motion control. At the same time, the model matching module also compares and analyzes the expected operation data with the actual operation data to evaluate the control effect and provide feedback to the control system for parameter adjustment or model optimization.

[0050] By matching the motion control model that best suits the current environmental conditions, the model matching module can significantly improve the control accuracy of the ROV robot. In the face of complex and changing underwater environments, the model matching module can flexibly select and adjust the motion control model, so that the ROV robot can better adapt to the operating requirements under different environmental conditions. By substituting real-time control commands and generating expected operating data, the model matching module provides an important reference and basis for the control system, which helps to optimize the control strategy and improve control efficiency.

[0051] In summary, the model matching module is an indispensable part of the ROV motion control system. The model matching module provides strong support for the precise control and efficient operation of the ROV robot through functions such as real-time environmental data matching, motion control model selection, and real-time control command substitution.

[0052] The model optimization module transmits the real-time control command to the ROV robot, collects the real-time operation data of the ROV robot, and compares the real-time operation data of the ROV robot with the expected operation data of the ROV robot. If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the motion control model corresponding to the real-time environment data is optimized using a genetic algorithm according to the error data between the real-time operation data of the ROV robot and the expected operation data of the ROV robot to obtain the motion control model corresponding to the optimized real-time environment data, and the real-time control command is processed using the motion control model corresponding to the optimized real-time environment data to generate an optimized ROV robot control command, and the optimized ROV robot control command is sent to the ROV robot.

[0053] The model optimization module is responsible for continuously optimizing the motion control model in the ROV (remotely operated vehicle) motion control system according to the actual operation conditions to ensure that the ROV robot can respond to control commands more accurately. The following is a detailed functional description of the model optimization module: The model optimization module first receives the real-time control command from the control system and transmits it to the ROV robot to perform the corresponding action. At the same time, the module collects the operation data of the ROV robot in real time through sensors or data interfaces, including key parameters such as position, speed, attitude, thrust, etc.

[0054] The collected real-time operation data of the ROV robot is compared with the expected operation data generated by the previous model matching module. The difference between the two is analyzed and the error data is calculated, which reflects the deviation between the actual operation and the expected operation.

[0055] When it is found that the real-time operation data of the ROV robot is inconsistent with the expected operation data, the model optimization module starts the optimization process. Using optimization techniques such as genetic algorithms, the motion control model corresponding to the real-time environmental data is iteratively optimized based on the error data. The genetic algorithm searches for the optimal solution in the model parameter space by simulating natural selection and genetic mechanisms to minimize the error. After multiple iterations, the optimized motion control model corresponding to the real-time environmental data is obtained, which can more accurately reflect the motion characteristics of the ROV robot in the current environment.

[0056] The optimized motion control model is used to process the real-time control commands and generate optimized ROV robot control commands. The optimized control commands take into account the difference between the actual operation and the expected operation, and can more accurately guide the movement of the ROV robot. Finally, the optimized ROV robot control commands are sent to the ROV robot for more precise control. The model optimization module also includes a feedback mechanism for evaluating the optimized control effect and feeding the evaluation results back to the control system for further adjustment or optimization.

[0057] Specifically, the PID-based ROV motion control system of the present invention, the data detection module includes: The specific steps of matching the real-time environment data set in the preset algorithm model knowledge base to obtain the motion control model corresponding to the real-time environment data include: Extract features from real-time environmental data sets to obtain water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features; The extracted features are combined into feature vectors according to the preset format and order. The feature vectors will be used for searching and matching in the algorithm model knowledge base. The feature vector is used to search in the algorithm model knowledge base to obtain the motion control model that best matches the real-time environment data. The motion control model that best matches the real-time environment data is used to calculate the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base; According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

[0058] In the ROV motion control system based on PID (proportional-integral-derivative), the data detection module matches the real-time environmental data set with the preset algorithm model knowledge base to find the motion control model that best suits the current environmental conditions.

[0059] The specific steps of model matching in the data detection module extract key features from the real-time environmental data set, which can reflect the impact of the environment on the motion control of the ROV robot. The extracted features include water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features.

[0060] The extracted features are combined into a feature vector in a preset format and order. Each element of the feature vector corresponds to a specific environmental feature, and its value represents the specific value or state of the feature in the current environment. The purpose of combining feature vectors is to facilitate searching and matching in the algorithm model knowledge base.

[0061] The combined feature vector is used to search in the algorithm model knowledge base. The algorithm model knowledge base is a database containing multiple motion control models and their corresponding feature vectors. The purpose of the search is to find the motion control model that best matches the current real-time environment data.

[0062] During the search process, the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base is calculated. The similarity can be obtained by calculating the Euclidean distance, cosine similarity or other similarity measurement methods between two vectors. The higher the similarity, the closer the current real-time environmental data is to the environmental conditions corresponding to the model.

[0063] According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

[0064] Specifically, the PID-based ROV motion control system of the present invention, the model matching module includes: Sending real-time control commands to the ROV robot through a communication link, the real-time control commands including target position commands, speed commands and attitude commands; After the ROV robot receives the control command, the internal control system of the ROV robot interprets the command and drives the corresponding actuator to move; In the process of the ROV robot executing the control command, the operating data of the ROV robot is collected in real time through the sensors installed on the ROV robot. The sensors on the ROV robot include position sensors, speed sensors and attitude sensors. The operating data of the ROV robot include actual position, actual speed, actual attitude, motor current and voltage; Recalculate the expected operation data of the ROV robot predicted by the motion control model based on real-time environmental data and control commands; The preprocessed real-time operation data of the ROV robot is compared with the expected operation data item by item, and the comparison results are output to the model optimization module.

[0065] In the ROV (remotely operated vehicle) motion control system based on PID (proportional-integral-derivative), the model matching module is not only responsible for matching the real-time environmental data to the most appropriate motion control model, but also involves the sending of control commands, the collection of real-time operation data, and the recalculation and comparison of expected operation data.

[0066] The real-time control command is sent to the ROV robot through a communication link (such as wireless or wired communication). The real-time control command usually includes a target position instruction, a speed instruction, and a posture instruction, which define the position, speed, and posture that the ROV robot should achieve.

[0067] After the ROV robot receives the control command, its internal control system interprets the command. According to the interpreted command, the ROV robot drives the corresponding actuators (such as thrusters, steering gears, etc.) to move.

[0068] When the ROV robot executes control commands, its operation data is collected in real time through sensors installed on the ROV robot. Sensors include position sensors (such as GPS, depth sensors), speed sensors (such as tachometers), attitude sensors (such as gyroscopes, accelerometers), and motor current and voltage sensors. The collected operation data includes the actual position, actual speed, actual attitude, motor current and voltage of the ROV robot.

[0069] Using the current real-time environment data and control commands, the expected operation data of the ROV robot is recalculated through the previously matched motion control model. The expected operation data is the state that the ROV robot should achieve according to the current environment and control requirements predicted by the model.

[0070] The collected real-time operation data of the ROV robot is preprocessed, such as filtering and denoising, to ensure the accuracy of the data. The preprocessed real-time operation data is compared item by item with the recalculated expected operation data. The comparison results include position error, speed error, attitude error, and motor working status.

[0071] The comparison results are output to the model optimization module, which uses optimization algorithms (such as genetic algorithms, particle swarm algorithms, etc.) to adjust and optimize the motion control model based on these comparison results to improve the accuracy of the model.

[0072] Specifically, the PID-based ROV motion control system of the present invention, the model optimization module includes: If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the specific steps of using a genetic algorithm to optimize the motion control model corresponding to the real-time environmental data and generate an optimized ROV robot control command include: Calculate error data, which includes position error, velocity error and attitude error; Taking reducing the error as the optimization goal, define the fitness function; Initialize the genetic algorithm and set the parameters of the genetic algorithm. Each individual represents a set of motion control model parameters. Use each individual in the population, i.e., the model parameter set, to simulate the real-time control command and obtain the simulated ROV robot operation data; Calculate the error between the simulated data and the real-time running data, and use it to evaluate the fitness of each individual; According to the fitness, excellent individuals are selected as parents, crossover operations are performed to generate offspring, and mutation operations are performed on the offspring. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirement is reached; Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data; Use the optimized motion control model to process the real-time control commands and obtain the expected operation data of the ROV robot; Based on the optimized expected operation data, an optimized ROV robot control command is generated, and the optimized ROV robot control command is sent to the ROV robot through a communication link to adjust the ROV robot motion state so that the ROV robot reaches the expected operation state.

[0073] In the ROV (remotely operated vehicle) motion control system based on PID (proportional-integral-differential), it is responsible for optimizing the motion control model through genetic algorithm and generating optimized control commands when the real-time operating data of the ROV robot is inconsistent with the expected operating data.

[0074] Calculate the error between the real-time operation data of the ROV robot and the expected operation data. The error includes position error, speed error and attitude error.

[0075] Reducing the error is taken as the optimization goal, and the fitness function is defined accordingly. The fitness function is a standard for evaluating the quality of each individual (i.e., the motion control model parameter set) in the genetic algorithm. The fitness function is inversely proportional to the error data, that is, the smaller the error, the higher the fitness.

[0076] Set the parameters of the genetic algorithm, such as population size, number of iterations, crossover probability, mutation probability, etc. Each individual represents a motion control model parameter set, which is the basis for genetic algorithm search and optimization. Use each individual (model parameter set) in the population to simulate the real-time control command and obtain the simulated ROV robot operation data. Calculate the error between the simulation data and the real-time operation data, and evaluate the fitness of each individual accordingly. Fitness evaluation is the basis for the genetic algorithm to select excellent individuals.

[0077] Selection, crossover and mutation: Select excellent individuals as parents based on fitness, and perform crossover operations to generate offspring.

[0078] Perform mutation operations on the offspring to increase the diversity of the population. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirements are reached.

[0079] Select the best individual: Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data. The best individual represents the motion control model parameters that can minimize the error under the current environmental conditions.

[0080] The real-time control command is processed using the optimized motion control model to obtain the expected operation data of the ROV robot. Based on the optimized expected operation data, an optimized ROV robot control command is generated. The optimized ROV robot control command is sent to the ROV robot through a communication link. After receiving the optimized control command, the ROV robot will adjust its motion state to more accurately achieve the expected operation state.

[0081] Second, see Figure 1 The present invention provides a PID-based ROV motion control method, which is applied to the PID-based ROV motion control system, comprising: Step S101, acquiring motion control system historical data, the motion control system historical data including target position historical data, actual position historical data, deviation historical data, control parameter historical data, environment historical data, sensor historical data, motion trajectory historical data and control instruction historical data; Step S102, extracting data features from the environmental data to obtain environmental data classification features, matching the environmental classification features with the motion control system historical data based on time information to obtain an environmental classification feature data set, dividing each type of environmental feature data in the environmental classification feature data set into a training set and a verification set, and using each type of environmental feature data divided into a training set and a verification set to train a neural network model to obtain a motion control model data set corresponding to the environmental classification features, and constructing the motion control model data set corresponding to the environmental classification features into an algorithm model knowledge base; Step S103, receiving real-time target location data and real-time water flow data, real-time wave data and real-time load data, acquiring target location natural environment data according to the real-time target location data, and aggregating the real-time water flow data, real-time wave data, real-time load data and target location natural environment data to obtain a real-time environment data set; Step S104, matching the real-time environment data set in a preset algorithm model knowledge base, obtaining a motion control model corresponding to the real-time environment data, receiving a real-time control command, substituting the real-time control command into the motion control model corresponding to the real-time environment data, and obtaining the expected operation data of the ROV robot; Step S105, transmit the real-time control command to the ROV robot, collect the real-time operation data of the ROV robot, compare the real-time operation data of the ROV robot with the expected operation data of the ROV robot, if the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, then use a genetic algorithm to optimize the motion control model corresponding to the real-time environment data according to the error data between the real-time operation data of the ROV robot and the expected operation data of the ROV robot, obtain the motion control model corresponding to the optimized real-time environment data, use the motion control model corresponding to the optimized real-time environment data to perform data processing on the real-time control command, generate an optimized ROV robot control command, and send the optimized ROV robot control command to the ROV robot.

[0082] Specifically, the PID-based ROV motion control method of the present invention, step S103, includes: The specific steps of matching the real-time environment data set in the preset algorithm model knowledge base to obtain the motion control model corresponding to the real-time environment data include: Extract features from real-time environmental data sets to obtain water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features; The extracted features are combined into feature vectors according to the preset format and order. The feature vectors will be used for searching and matching in the algorithm model knowledge base. The feature vector is used to search in the algorithm model knowledge base to obtain the motion control model that best matches the real-time environment data. The motion control model that best matches the real-time environment data is used to calculate the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base; According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

[0083] Specifically, the PID-based ROV motion control method of the present invention, step S104, includes: Sending real-time control commands to the ROV robot through a communication link, the real-time control commands including target position commands, speed commands and attitude commands; After the ROV robot receives the control command, the internal control system of the ROV robot interprets the command and drives the corresponding actuator to move; In the process of the ROV robot executing the control command, the operating data of the ROV robot is collected in real time through the sensors installed on the ROV robot. The sensors on the ROV robot include position sensors, speed sensors and attitude sensors. The operating data of the ROV robot include actual position, actual speed, actual attitude, motor current and voltage; Recalculate the expected operation data of the ROV robot predicted by the motion control model based on real-time environmental data and control commands; The preprocessed real-time operation data of the ROV robot is compared with the expected operation data item by item, and the comparison results are output to the model optimization module.

[0084] Specifically, the PID-based ROV motion control method of the present invention, step S105, includes: If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the specific steps of using a genetic algorithm to optimize the motion control model corresponding to the real-time environmental data and generate an optimized ROV robot control command include: Calculate error data, which includes position error, velocity error and attitude error; Taking reducing the error as the optimization goal, define the fitness function; Initialize the genetic algorithm and set the parameters of the genetic algorithm. Each individual represents a set of motion control model parameters. Use each individual in the population, i.e., the model parameter set, to simulate the real-time control command and obtain the simulated ROV robot operation data; Calculate the error between the simulated data and the real-time running data, and use it to evaluate the fitness of each individual; According to the fitness, excellent individuals are selected as parents, crossover operations are performed to generate offspring, and mutation operations are performed on the offspring. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirement is reached; Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data; Use the optimized motion control model to process the real-time control commands and obtain the expected operation data of the ROV robot; Based on the optimized expected operation data, an optimized ROV robot control command is generated, and the optimized ROV robot control command is sent to the ROV robot through a communication link to adjust the ROV robot motion state so that the ROV robot reaches the expected operation state.

[0085] The present invention solves the problem that the existing PID control algorithm cannot provide accurate control effect when facing a complex system such as ROV through the following technical solutions: Acquire motion control system historical data, including target position historical data, actual position historical data, deviation historical data, control parameter historical data, environment historical data, sensor historical data, motion trajectory historical data, and control instruction historical data.

[0086] The environmental data is feature extracted to obtain the environmental data classification features, and these features are matched with the motion control system historical data based on time information to construct an environmental classification feature data set. These data sets are used to train the neural network model to obtain the motion control model data set corresponding to different environmental classification features, and finally to construct an algorithm model knowledge base.

[0087] Receive real-time target location data, real-time water flow data, real-time wave data, real-time load data, etc., obtain the natural environment data of the target location, and summarize these data into a real-time environment data set. Match the real-time environment data set in the preset algorithm model knowledge base, and select the motion control model that best matches the current real-time environment data through steps such as feature extraction, feature vector combination, search and similarity calculation.

[0088] Receive real-time control commands (such as target position commands, speed commands, attitude commands, etc.), substitute them into the motion control model corresponding to the real-time environmental data, and calculate the expected operation data of the ROV robot. Transmit the real-time control commands to the ROV robot, and collect the real-time operation data of the ROV robot. Compare the real-time operation data of the ROV robot with the expected operation data. If the two are inconsistent, use the genetic algorithm to optimize the motion control model corresponding to the real-time environmental data based on the error data (such as position error, speed error, attitude error, etc.). The genetic algorithm iteratively optimizes the model parameters through operations such as selection, crossover, and mutation until the preset number of iterations or error range requirements are reached, thereby obtaining the optimized motion control model.

[0089] The optimized motion control model is used to process the real-time control command data and generate the optimized ROV robot control command. The optimized control command is sent to the ROV robot to adjust its motion state to achieve the expected operating state.

[0090] Through the above technical scheme, the present invention can dynamically select and optimize the motion control model when the ROV faces a complex and changeable underwater environment, thereby providing a more accurate control effect and solving the shortcomings of the traditional PID control algorithm when facing nonlinear and time-varying systems.

Claims

1. A PID-based ROV motion control system, characterized in that: include: A data acquisition module is used to acquire the historical data of the motion control system, which includes the target position historical data, the actual position historical data, the deviation historical data, the control parameter historical data, the environment historical data, the sensor historical data, the motion trajectory historical data and the control instruction historical data; A data processing module is used to extract data features from environmental data to obtain environmental data classification features, match the environmental classification features with historical data of the motion control system based on time information to obtain an environmental classification feature data set, divide each type of environmental feature data in the environmental classification feature data set into a training set and a verification set, and use each type of environmental feature data divided into a training set and a verification set to train a neural network model to obtain a motion control model data set corresponding to the environmental classification features, and construct the motion control model data set corresponding to the environmental classification features into an algorithm model knowledge base; A data detection module is used to receive real-time target location data and real-time water flow data, real-time wave data and real-time load data, obtain target location natural environment data according to the real-time target location data, and aggregate the real-time water flow data, real-time wave data, real-time load data and target location natural environment data to obtain a real-time environment data set; The model matching module is used to match the real-time environment data set in the preset algorithm model knowledge base, obtain the motion control model corresponding to the real-time environment data, receive the real-time control command, substitute the real-time control command into the motion control model corresponding to the real-time environment data, and obtain the expected operation data of the ROV robot; The model optimization module is used to transmit the real-time control command to the ROV robot, collect the real-time operation data of the ROV robot, and compare the real-time operation data of the ROV robot with the expected operation data of the ROV robot. If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the motion control model corresponding to the real-time environment data is optimized using a genetic algorithm according to the error data between the real-time operation data of the ROV robot and the expected operation data of the ROV robot to obtain the motion control model corresponding to the optimized real-time environment data, and the real-time control command is processed using the motion control model corresponding to the optimized real-time environment data to generate an optimized ROV robot control command, and the optimized ROV robot control command is sent to the ROV robot.

2. The ROV motion control system based on PID according to claim 1, characterized in that: The data detection module is further used for: Extract features from real-time environmental data sets to obtain water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features; The extracted features are combined into feature vectors according to the preset format and order. The feature vectors will be used for searching and matching in the algorithm model knowledge base. The feature vector is used to search in the algorithm model knowledge base to obtain the motion control model that best matches the real-time environment data. The motion control model that best matches the real-time environment data is used to calculate the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base; According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

3. The PID-based ROV motion control system according to claim 1, characterized in that: The model matching module is also used for: Sending real-time control commands to the ROV robot through a communication link, the real-time control commands including target position commands, speed commands and attitude commands; After the ROV robot receives the control command, the internal control system of the ROV robot interprets the command and drives the corresponding actuator to move; In the process of the ROV robot executing the control command, the operating data of the ROV robot is collected in real time through the sensors installed on the ROV robot. The sensors on the ROV robot include position sensors, speed sensors and attitude sensors. The operating data of the ROV robot include actual position, actual speed, actual attitude, motor current and voltage; Recalculate the expected operation data of the ROV robot predicted by the motion control model based on real-time environmental data and control commands; The preprocessed real-time operation data of the ROV robot is compared with the expected operation data item by item, and the comparison results are output to the model optimization module.

4. The PID-based ROV motion control system according to claim 3, characterized in that: The model optimization module is also used for: Calculate error data, which includes position error, velocity error and attitude error; Taking reducing the error as the optimization goal, define the fitness function; Initialize the genetic algorithm and set the parameters of the genetic algorithm. Each individual represents a set of motion control model parameters. Use each individual in the population, i.e., the model parameter set, to simulate the real-time control command and obtain the simulated ROV robot operation data; Calculate the error between the simulated data and the real-time running data, and use it to evaluate the fitness of each individual; According to the fitness, excellent individuals are selected as parents, crossover operations are performed to generate offspring, and mutation operations are performed on the offspring. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirement is reached; Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data; Use the optimized motion control model to process the real-time control commands and obtain the expected operation data of the ROV robot; Based on the optimized expected operation data, an optimized ROV robot control command is generated, and the optimized ROV robot control command is sent to the ROV robot through a communication link to adjust the ROV robot motion state so that the ROV robot reaches the expected operation state.

5. A PID-based ROV motion control method, applied to the PID-based ROV motion control system according to any one of claims 1 to 4, characterized in that: include: Step S101, acquiring motion control system historical data, the motion control system historical data including target position historical data, actual position historical data, deviation historical data, control parameter historical data, environment historical data, sensor historical data, motion trajectory historical data and control instruction historical data; Step S102, extracting data features from the environmental data to obtain environmental data classification features, matching the environmental classification features with the motion control system historical data based on time information to obtain an environmental classification feature data set, dividing each type of environmental feature data in the environmental classification feature data set into a training set and a verification set, and using each type of environmental feature data divided into a training set and a verification set to train a neural network model to obtain a motion control model data set corresponding to the environmental classification features, and constructing the motion control model data set corresponding to the environmental classification features into an algorithm model knowledge base; Step S103, receiving real-time target location data and real-time water flow data, real-time wave data and real-time load data, acquiring target location natural environment data according to the real-time target location data, and aggregating the real-time water flow data, real-time wave data, real-time load data and target location natural environment data to obtain a real-time environment data set; Step S104, matching the real-time environment data set in a preset algorithm model knowledge base, obtaining a motion control model corresponding to the real-time environment data, receiving a real-time control command, substituting the real-time control command into the motion control model corresponding to the real-time environment data, and obtaining the expected operation data of the ROV robot; Step S105, transmit the real-time control command to the ROV robot, collect the real-time operation data of the ROV robot, compare the real-time operation data of the ROV robot with the expected operation data of the ROV robot, if the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, then use a genetic algorithm to optimize the motion control model corresponding to the real-time environment data according to the error data between the real-time operation data of the ROV robot and the expected operation data of the ROV robot, obtain the motion control model corresponding to the optimized real-time environment data, use the motion control model corresponding to the optimized real-time environment data to perform data processing on the real-time control command, generate an optimized ROV robot control command, and send the optimized ROV robot control command to the ROV robot.

6. The ROV motion control method based on PID as claimed in claim 5, characterized in that: The step S103 includes: The specific steps of matching the real-time environment data set in the preset algorithm model knowledge base to obtain the motion control model corresponding to the real-time environment data include: Extract features from real-time environmental data sets to obtain water velocity features, wave height features, load weight features, water temperature features, salinity features, and seabed topography features; The extracted features are combined into feature vectors according to the preset format and order. The feature vectors will be used for searching and matching in the algorithm model knowledge base. The feature vector is used to search in the algorithm model knowledge base to obtain the motion control model that best matches the real-time environment data. The motion control model that best matches the real-time environment data is used to calculate the similarity between the feature vector and the feature vectors corresponding to each model in the knowledge base; According to the calculation results of the similarity, the motion control model with the highest similarity is selected as the best matching model corresponding to the current real-time environment data.

7. The ROV motion control method based on PID as claimed in claim 5, characterized in that: The step S104 includes: Sending real-time control commands to the ROV robot through a communication link, the real-time control commands including target position commands, speed commands and attitude commands; After the ROV robot receives the control command, the internal control system of the ROV robot interprets the command and drives the corresponding actuator to move; In the process of the ROV robot executing the control command, the operating data of the ROV robot is collected in real time through the sensors installed on the ROV robot. The sensors on the ROV robot include position sensors, speed sensors and attitude sensors. The operating data of the ROV robot include actual position, actual speed, actual attitude, motor current and voltage; Recalculate the expected operation data of the ROV robot predicted by the motion control model based on real-time environmental data and control commands; The preprocessed real-time operation data of the ROV robot is compared with the expected operation data item by item, and the comparison results are output to the model optimization module.

8. The ROV motion control method based on PID according to claim 7, characterized in that: The step S105 includes: If the real-time operation data of the ROV robot is inconsistent with the expected operation data of the ROV robot, the specific steps of using a genetic algorithm to optimize the motion control model corresponding to the real-time environmental data and generate an optimized ROV robot control command include: Calculate error data, which includes position error, velocity error and attitude error; Taking reducing the error as the optimization goal, define the fitness function; Initialize the genetic algorithm and set the parameters of the genetic algorithm. Each individual represents a set of motion control model parameters. Use each individual in the population, i.e., the model parameter set, to simulate the real-time control command and obtain the simulated ROV robot operation data; Calculate the error between the simulated data and the real-time running data, and use it to evaluate the fitness of each individual; According to the fitness, excellent individuals are selected as parents, crossover operations are performed to generate offspring, and mutation operations are performed on the offspring. Repeat the steps of fitness evaluation, selection, crossover, and mutation until the preset number of iterations or error range requirement is reached; Select the individual with the highest fitness from the final population as the motion control model parameter set corresponding to the optimized real-time environmental data; Use the optimized motion control model to process the real-time control commands and obtain the expected operation data of the ROV robot; Based on the optimized expected operation data, an optimized ROV robot control command is generated, and the optimized ROV robot control command is sent to the ROV robot through a communication link to adjust the ROV robot motion state so that the ROV robot reaches the expected operation state.

Images