Dynamic environment sensing and navigation device and method for underwater robot

Abstract

The invention discloses an underwater robot dynamic environment sensing and navigation device and method, and relates to the technical field of underwater robos.The device comprises an underwater robot, a sensor system, a data processing system, a control system, a navigation system and a power system, and the sensor system, the data processing system, the control system, the navigation system and the power system are arranged on the underwater robot; the surrounding environment can be comprehensively sensed in real time, and the optimal path is calculated through the path planning module according to the real-time environment data; performing obstacle detection according to the environment image to obtain an obstacle detection result; when an obstacle is detected, the control system sends out an obstacle avoidance decision instruction, after the path planning module receives the obstacle avoidance decision instruction, an adaptive algorithm is used for adjusting the optimal path according to the information of the obstacle to be detected, the adjusted path is obtained, and autonomous navigation for a complex underwater environment and autonomous navigation are achieved. And the autonomous navigation and operation capabilities of the underwater robot are obviously improved.

Description

Technical Field

[0001] The present application relates to the technical field of underwater robots, and in particular to a dynamic environment perception and navigation device and method for underwater robots. Background Art

[0002] Current underwater detection technology faces challenges such as complex environments and incomplete data. Existing underwater robots, such as BlueROV2, OceanOne, Seaglider, and Remus 600, have their own advantages, but they still have insufficient autonomy. Therefore, there is an urgent need for an underwater robot dynamic environment perception and navigation device that can perceive unknown environments and navigate autonomously. Summary of the invention

[0003] The purpose of this application is to provide an underwater robot dynamic environment perception and navigation device and method to achieve real-time perception and autonomous navigation of complex underwater environments.

[0004] To achieve the above objectives, this application provides the following solutions:

[0005] In a first aspect, the present application provides an underwater robot dynamic environment perception and navigation device, including an underwater robot and a sensor system, a data processing system, a control system, a navigation system and a power system arranged on the underwater robot;

[0006] The sensor system includes a sonar sensor, a camera, a lidar, a Doppler velocity meter, a depth sensor, a temperature sensor, a water quality sensor and an inertial measurement unit;

[0007] The sonar sensor is used to scan the underwater environment and detect long-distance obstacles to obtain long-distance information; the long distance is a distance that falls within a first set threshold range;

[0008] The camera is used to collect environmental images of the underwater environment;

[0009] The laser radar is used to scan the underwater environment and detect short-range obstacles to obtain short-range information; the short-range is a distance falling within a second set threshold range; the long-range information, the short-range information and the environment image constitute the underwater obstacle information;

[0010] The Doppler velocity measuring instrument is used to measure the speed and position of the underwater robot; the depth sensor is used to measure the depth of the underwater robot in the underwater environment; the temperature sensor is used to measure the temperature of the underwater environment; the water quality sensor is used to measure the water quality information of the underwater environment; the inertial measurement unit is used to measure the motion state and direction of the underwater robot; the speed and position of the underwater robot, the depth of the underwater robot in the underwater environment, the temperature of the underwater environment, the water quality information of the underwater environment, and the motion state and direction of the underwater robot constitute the surrounding environment data;

[0011] The data processing system is used to pre-process the surrounding environment data and underwater obstacle information respectively to obtain real-time environment data and obstacle information to be detected;

[0012] The control system includes a central processing unit for controlling the navigation system and the power system;

[0013] The navigation system includes a path planning module and a dynamic obstacle avoidance module; the path planning module is used to calculate the optimal path according to the real-time environmental data; the dynamic obstacle avoidance module is used to: detect obstacles according to the environmental image to obtain obstacle detection results; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after receiving the obstacle avoidance decision instruction, the path planning module uses an adaptive algorithm to adjust the optimal path according to the obstacle information to be detected to obtain an adjusted path;

[0014] The power system is used to provide the underwater robot with power for movement and steering, so as to control the movement of the underwater robot according to the path output by the path planning module.

[0015] Optionally, the underwater robot dynamic environment perception and navigation device further comprises a front flow field information collection system; the front flow field information collection system is used to: collect the flow field information of the underwater environment in real time to obtain initial flow field information; the flow field information includes water flow velocity, direction and turbulence characteristics;

[0016] The data processing system pre-processes the initial flow field information to obtain the flow field information to be detected;

[0017] The flow field information to be detected is updated into the flow field model. If it is detected that the difference between the flow field information to be detected and the set flow field threshold is greater than the set flow field threshold, the output parameters of the navigation system are adjusted.

[0018] Optionally, the underwater robot dynamic environment perception and navigation device also includes a shock absorption system; the shock absorption system is installed at various key positions of the underwater robot and is used to: measure the external force applied to the underwater robot through sensors to obtain the external force; determine whether the external force is greater than a set external force threshold, and if so, calculate new damping parameters.

[0019] Optionally, the underwater robot dynamic environment perception and navigation device also includes a pressure detection system; the pressure detection system includes a pressure processing unit and a plurality of pressure detection points; the pressure detection points are installed at key parts of the underwater robot shell and interior; the pressure detection points are used to measure the real-time pressure of key parts of the underwater robot; the pressure processing unit is used to: for each key part, determine whether the real-time pressure of the key part is greater than the pressure safety limit corresponding to the key part, and if so, generate a safety alarm.

[0020] Optionally, in terms of calculating the optimal path according to the real-time environmental data, the path planning module is used to: calculate the optimal path according to the real-time environmental data using an A* algorithm and a Dijkstra algorithm.

[0021] Optionally, the power system includes a propeller and a steering device; the propeller is used to provide driving force through a brushless motor; and the steering device is used to control the moving direction of the underwater robot through a servo control system.

[0022] Optionally, the power system also includes an intelligent power management system; the intelligent power management system is used to measure the real-time battery status of the battery; the real-time battery status includes voltage, current and temperature; determine whether the real-time battery status exceeds the battery safety limit; if so, generate a safety alarm.

[0023] Optionally, the underwater robot dynamic environment perception and navigation device also includes a communication system; the sensor system, data processing system, control system, navigation system and power system are connected through the communication system.

[0024] In a second aspect, the present application provides a method for underwater robot dynamic environment perception and navigation, comprising:

[0025] Detecting the underwater environment in which the underwater robot is located to obtain surrounding environment data; the surrounding environment data includes the speed and position of the underwater robot, the depth of the underwater robot in the underwater environment, the temperature of the underwater environment, the water quality information of the underwater environment, and the movement state and direction of the underwater robot;

[0026] Scanning the underwater environment and detecting obstacles to obtain long-distance information and short-distance information; the long-distance is a distance falling within a first set threshold range; the short-distance is a distance falling within a second set threshold range;

[0027] Collecting an environmental image of an underwater environment; the long-distance information, the short-distance information and the environmental image constitute underwater obstacle information;

[0028] Preprocess the surrounding environment data and underwater obstacle information respectively to obtain real-time environment data and obstacle information to be detected;

[0029] The optimal path is calculated based on real-time environmental data through the path planning module;

[0030] Obstacle detection is performed according to the environment image to obtain an obstacle detection result; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after the path planning module receives the obstacle avoidance decision instruction, the path planning module uses an adaptive algorithm to adjust the optimal path according to the obstacle information to be detected to obtain an adjusted path;

[0031] The underwater robot is controlled to move according to the path output by the path planning module.

[0032] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0033] The present application provides a dynamic environment perception and navigation device and method for an underwater robot. Through the sensors in the sensor system, the device can fully and real-time perceive the surrounding environment, and calculate the optimal path according to the real-time environmental data through the path planning module; obstacle detection is performed according to the environmental image to obtain the obstacle detection result; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after the path planning module receives the obstacle avoidance decision instruction, it uses an adaptive algorithm to adjust the optimal path according to the information of the obstacle to be detected, and obtains the adjusted path, thereby realizing autonomous navigation in complex underwater environments and significantly improving the autonomous navigation and operation capabilities of the underwater robot. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 A schematic diagram of functional modules of a dynamic environment perception and navigation device for an underwater robot provided in one embodiment of the present application;

[0036] Figure 2 A flowchart of a method for dynamic environment perception and navigation of an underwater robot provided in one embodiment of the present application. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0038] Existing underwater robots also have problems such as limited data processing capabilities and imperfect energy management.

[0039] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0040] In an exemplary embodiment, Figure 1 As shown, a dynamic environment perception and navigation device for an underwater robot is provided, including an underwater robot and a sensor system, a data processing system, a control system, a navigation system, a power system, a front flow field information collection system and a communication system arranged on the underwater robot; the sensor system, the data processing system, the control system, the navigation system, the power system and the front flow field information collection system are connected through the communication system.

[0041] The sensor system can provide reliable environmental data under different water depths and water quality conditions, collect surrounding environment data, and transmit the data to the data processing module for preliminary processing. The sensor system includes sonar sensor (SonarSensor), camera, lidar, Doppler velocity meter (Doppler Velocity Log, DVL), depth sensor (Depth Sensor), temperature sensor, water quality sensor and inertial measurement unit (Inertial MeasurementUnit, IMU). Each sensor is installed in different parts of the underwater robot to fully perceive the surrounding environment of the underwater robot.

[0042] The sonar sensor is used to scan the underwater environment and detect long-distance obstacles to obtain long-distance information; the long distance is the distance that falls within the first set threshold range. The long-distance information includes long-distance obstacle information and terrain changes in the underwater environment. The sonar sensor uses multi-beam technology and can provide high-resolution terrain data in complex terrain. The operating frequency is 200kHz and the maximum detection range is 300 meters.

[0043] The camera is used to collect images of the underwater environment and obtain visual images of the environment. It uses high-sensitivity low-light imaging technology to enable effective visual perception in low-light environments. It includes parameters such as aperture, focal length, resolution, and field of view.

[0044] Laser radar is used to scan the underwater environment and detect close-range obstacles to obtain close-range information; the close distance is the distance that falls within the second set threshold range; it uses 360-degree scanning technology to ensure all-round close-range object detection, with a wavelength of 905nm, a scanning frequency of 10Hz, and a ranging accuracy of ±2cm, which can accurately measure close-range objects and obstacles.

[0045] Long-range information, short-range information and environmental images constitute underwater obstacle information.

[0046] The Doppler velocity meter is used to measure the speed and position of the underwater robot. The speed is calculated by reflecting sound waves through the Doppler effect. The measured speed and position of the underwater robot are used for navigation and motion control.

[0047] Depth sensors are used to measure the depth of underwater robots in underwater environments; pressure sensors are used to measure underwater depth; they are used for depth measurement and diving depth control.

[0048] The temperature sensor is used to measure the temperature of the underwater environment; a thermocouple or RTD sensor is used to measure the water temperature. The measured temperature of the underwater environment is used for environmental monitoring and temperature regulation.

[0049] The water quality sensor is used to measure the water quality information of the underwater environment. Specifically, a multi-parameter water quality sensor is used to measure the water quality information of the water; the water quality information includes pH value, dissolved oxygen, turbidity, etc. The measured water quality information is used for environmental monitoring and data collection.

[0050] The inertial measurement unit, which includes an accelerometer, a gyroscope and a magnetometer, is used to measure the motion state and direction of the underwater robot; the measured motion state and direction of the underwater robot are used for navigation and attitude control.

[0051] The speed and position of the underwater robot, the depth of the underwater robot in the underwater environment, the temperature of the underwater environment, the water quality information of the underwater environment, and the movement state and direction of the underwater robot constitute the surrounding environment data.

[0052] The sensor system also includes a sensor repeater and a data storage module, which includes a high-reliability data storage submodule and a sensor repeater. The built-in black box data storage submodule can record all sensor data and system status information of the underwater robot during operation. When an abnormal situation occurs, it can provide detailed data records to help analyze and restore the state before the accident. The sensor repeater integrates the data of various sensors and performs preliminary processing to ensure the integrity and real-time nature of the data.

[0053] Specifically, the sensor repeater and data storage algorithm are as follows: initialize the data storage system to store sensor data, collect data from each sensor and perform preliminary processing, and store the processed data in the data storage system. When the storage system capacity is close to full, perform data backup and clean up old data. Continuously monitor the system status, detect faults and take corresponding measures. Transmit the processed data to the base station for further analysis.

[0054] The sensor repeater performs preliminary processing on the data of each sensor, focusing on data collection, integration and real-time transmission. The sensor repeater integrates the data of various sensors and performs preliminary integration processing to ensure the integrity and real-time nature of the data. It is also used to store data in a high-reliability storage submodule (such as a black box) for analysis in the event of an abnormal situation.

[0055] The data processing system is used to pre-process the surrounding environment data and underwater obstacle information respectively to obtain real-time environment data and obstacle information to be detected. Specifically: the data collected by the sensor system is preliminarily processed and filtered, and real-time environment data and obstacle information to be detected are provided to the control system and navigation system to assist in path decision-making.

[0056] Preprocessing of the data processing system: mainly focuses on in-depth analysis and optimization of the collected sensor data. The preprocessing process includes steps such as filtering, noise reduction, data verification and format conversion, aiming to extract useful information and generate real-time environmental data and obstacle information to be detected suitable for control systems and navigation systems.

[0057] The data processing system includes a data fusion module and a feature extraction module. In the data preprocessing stage, the data processing system fuses the multi-source data from the sensor system. Specifically, it integrates the surrounding environment data (such as speed, position, depth, temperature, water quality information, motion state and direction) with the underwater obstacle information (such as long-distance information, short-distance information and environmental images).

[0058] The purpose of data fusion in the data fusion module is to combine the advantages of different sensors, eliminate the blind spots or errors that may exist in a single sensor, and improve the accuracy and completeness of environmental perception. For example, the long-range detection capability of the sonar sensor is combined with the high-precision close-range measurement of the lidar.

[0059] After completing data fusion, the data processing system performs feature extraction on the fused data. The feature extraction process includes extracting key features useful for navigation and obstacle avoidance from environmental data and obstacle information.

[0060] For example, the shape, size, position and other features of obstacles are extracted from environmental images; the changing features of water flow speed and direction are extracted from flow field information; and the posture change features of the robot are extracted from the data of the inertial measurement unit.

[0061] The real-time environmental data and obstacle information to be detected after being processed by the data fusion module and feature extraction module are passed to the path planning module in the navigation system. The path planning module uses these high-quality feature data in combination with algorithms (such as the A* algorithm and the Dijkstra algorithm) to calculate the optimal path. Data fusion and feature extraction ensure that the path planning module makes decisions based on accurate and comprehensive environmental information. During the dynamic obstacle avoidance process, the extracted features are used for obstacle detection and recognition by deep neural networks. When an obstacle is detected, the system uses previously extracted obstacle features (such as location, motion trend) and environmental features to generate obstacle avoidance strategies and adjust the path.

[0062] When processing the data collected by the sensor, the specific process is as follows: initialize an empty data storage structure to store the processed data; traverse each data point and filter it. If the data point is valid after filtering, it is added to the processing data structure. Return all processed and verified data.

[0063] The control system comprehensively processes the data provided by the data processing system, makes decisions and controls, manages and coordinates the operations of the navigation system and the power system, and transmits data and receives commands with external devices through the communication system; the control system includes a central processing unit and several data processing modules, which are used to control the navigation system and the power system.

[0064] Central Processing Unit (CPU): It uses a high-performance processor that can process large amounts of data in real time. The main frequency is 3.2GHz, the number of cores is 8, and the cache size is 16MB.

[0065] Data processing module: adopts parallel computing architecture and integrates GPU acceleration to improve data processing speed. The module has built-in deep learning algorithm and can perform real-time analysis of environmental data.

[0066] It should be noted that the control system adopts a distributed computing architecture, and the central processing unit works together with multiple data processing modules to efficiently process large amounts of data from various sensors. The control system also includes an adaptive algorithm module that can adjust the navigation strategy based on real-time data to ensure the efficient operation of the underwater robot in complex environments. Adaptive algorithm: Based on deep learning and reinforcement learning technology, it can self-adjust and optimize in different environments.

[0067] The navigation system performs path planning and dynamic obstacle avoidance based on the real-time data provided by the data processing module, controls the power system, and ensures that the underwater robot moves along the optimal path. It includes a path planning module and a dynamic obstacle avoidance module.

[0068] The path planning module is used to calculate the optimal path according to the real-time environment data using the A* algorithm and the Dijkstra algorithm, and specifically includes the following steps 101 to 110:

[0069] Step 101: Initialize the open list and closed list required for path planning, add the start node to the open list, and start path planning; the node is a location point in the underwater environment.

[0070] Step 102: When the open list is not empty, execute the following steps 103 to 109.

[0071] Step 103: Get the node with the lowest cost from the open list.

[0072] In the A* algorithm, the total cost of a node is calculated as: f(n) = g(n) + h(n);

[0073] Among them, f(n) is the total cost, which is used to compare the node priority in the open list; g(n) is the current cost, that is, the actual path cost (actual cost) from the starting point to the current node n; h(n) is the heuristic cost, that is, the estimated cost from the current node n to the target node.

[0074] The calculation process of g(n) is: the path cost from the starting point to node n. Assuming that the cost from the starting point to node a is g(a), the cost to reach node n is: g(n) = g(a) + cost(a,n); where cost(a,n) is the cost of the edge from node a to node n.

[0075] The heuristic cost h(n) is the estimated cost from node n to the target node. It is usually estimated by geometric distance, and the Euclidean distance algorithm and the Manhattan distance algorithm can be used:

[0076] When using Euclidean distance (i.e. straight-line distance), it can be expressed as follows:

[0077] h(n)=(x goal -x n ) 2 +(y goal -y n ) 2

[0078] When using Manhattan distance (suitable for grid maps), it can be expressed as:

[0079] h(n)=|xgoal -x n ∣+∣y goal -y n ∣

[0080] Among them, x goal and goal are the horizontal and vertical coordinates of the target node, respectively, n and n are the horizontal and vertical coordinates of node n respectively.

[0081] In the Dijkstra algorithm, heuristic estimation is not used when calculating the cost of a node. The cost of a node is only the actual cost g(n) from the starting point to the current node. The total cost of the node is calculated as follows: f(n) = g(n), where g(n) is the actual cost from the starting point to node n, and the calculation method is the same as the A* algorithm.

[0082] Step 104: Determine whether the target node has been reached: If the current node is the target node, return the path, remove the current node from the open list, and add the current node to the closed list.

[0083] Step 105: Traverse the neighboring nodes of the current node: perform the following steps 106 to 109 for each neighboring node (a node whose distance from the current node is less than a set distance):

[0084] Step 106: If the neighbor node is in the closed list, skip the neighbor node.

[0085] Step 107: Calculate the temporary cost from the current node to the neighboring node.

[0086] Step 108: If the neighbor node is not in the open list, or the temporary cost from the current node to the neighbor node is less than the current cost, the cost and parent node information of the neighbor node are updated.

[0087] Step 109: If the neighbor node is not in the open list, add it to the open list.

[0088] Step 110: If the open list is empty and the target node is not found, return failure.

[0089] The dynamic obstacle avoidance module is used to: use a deep neural network to detect obstacles based on the environment image collected by the camera, obtain obstacle detection results, and thus detect obstacles in the underwater environment; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after receiving the obstacle avoidance decision instruction, the path planning module uses an adaptive algorithm to adjust the optimal path according to the obstacle information to be detected, and obtains an adjusted path, which can quickly adjust the path when an obstacle is detected to avoid collision. In this embodiment, the deep neural network can use a convolutional neural network, and the dynamic obstacle avoidance based on the deep neural network can improve the accuracy and real-time performance of obstacle detection.

[0090] Once the dynamic obstacle avoidance module detects an obstacle, the central processing unit (CPU) of the control system will generate obstacle avoidance decision instructions based on the preset strategy and real-time environmental data. This process ensures that the robot can respond to environmental changes in a timely manner.

[0091] After receiving the obstacle avoidance decision instruction, the path planning module will call the adaptive algorithm to recalculate the current optimal path. The adaptive algorithm may combine real-time environmental data and obstacle information, using algorithms such as A*, Dijkstra, or other optimization algorithms to generate a new path to bypass obstacles.

[0092] Environmental adaptability algorithm steps: Initialize the parameters of the environmental adaptability algorithm, including environmental data and sensor feedback; collect real-time data of the surrounding environment, including temperature, pressure, flow field and other information; process the data fed back by the sensor to generate an environmental model; detect changes in environmental data to determine whether system parameters need to be adjusted; adjust the parameters of the control system and navigation system according to environmental changes to adapt to new environmental conditions; apply the adjusted parameters to the underwater robot system to ensure its normal operation in the new environment; continuously monitor the performance of the underwater robot in the adjusted environment to ensure its stability and efficiency; update the parameters of the environmental adaptability algorithm according to actual operation conditions to improve future adaptability.

[0093] The adaptive algorithm not only takes into account the currently detected obstacle information, but also dynamically adjusts the path based on factors such as the robot's motion state and environmental changes.

[0094] The control system plays a coordinating role in the entire process (dynamic path planning, dynamic obstacle avoidance and environmental adaptability adjustment), ensuring the information transmission and execution between sensor data, obstacle detection results, path planning modules and power systems.

[0095] This embodiment also uses reinforcement learning technology and applies it to dynamic path planning, dynamic obstacle avoidance and environmental adaptability adjustment scenarios to enhance the autonomous navigation and obstacle avoidance capabilities of the underwater robot. Specifically, reinforcement learning is used for adaptive algorithms in control systems and navigation systems, enabling the robot to learn and adjust its behavior strategy based on environmental feedback.

[0096] 1. Reinforcement learning technology is applied to dynamic path planning:

[0097] Problem: The underwater environment is complex and changeable, with uncertain ocean currents, obstacles and other environmental factors. Traditional preset paths may not be able to adapt to real-time changes.

[0098] Reinforcement learning solution: The robot uses reinforcement learning algorithms to learn the optimal path planning strategy through trial and error in the process of interacting with the environment. The robot takes the state of the environment as input (such as current position, speed, position of surrounding obstacles, etc.), takes actions (such as adjusting heading or speed), and receives rewards or penalties based on the results of the actions.

[0099] For example, when the robot finds that a certain path consumes more energy or has obstacles, the reinforcement learning algorithm will adjust its strategy to choose a path with lower energy consumption and less risk.

[0100] 2. Reinforcement learning technology is applied to dynamic obstacle avoidance:

[0101] Problem: There are dynamic obstacles underwater, such as moving marine life and floating objects, and traditional obstacle avoidance algorithms may not be able to respond in time.

[0102] Reinforcement learning solution: Through reinforcement learning, the robot can learn the best obstacle avoidance strategy in different situations. After detecting an obstacle, the robot takes different avoidance actions and obtains feedback based on the obstacle avoidance effect (such as whether it is successfully avoided, how much time it takes, etc.).

[0103] For example, after avoiding moving obstacles many times, the robot learned to predict the movement trajectory of the obstacles, so as to adjust its course in advance and avoid collisions.

[0104] 3. Reinforcement learning technology is applied to environmental adaptability adjustment:

[0105] Problem: Parameters of the underwater environment, such as temperature, pressure, and water quality, may affect the performance of the robot’s sensors and actuators.

[0106] Reinforcement learning solution: The robot uses reinforcement learning algorithms to adjust its control parameters according to real-time environmental parameters. For example, when it detects that the water temperature drops and the sensor sensitivity decreases, the robot can adjust the sensor's working mode or rely on data from other sensors.

[0107] For example, in the deep-sea high-pressure environment, the robot has learned to adjust the output power of the thrusters to overcome the impact of water pressure on its motion performance.

[0108] 4. Reinforcement learning technology is applied to energy efficiency optimization:

[0109] Problem: Underwater robots have limited battery capacity and need to conserve energy as much as possible while performing their missions.

[0110] Reinforcement learning solution: Through reinforcement learning, the robot can learn energy optimization strategies for different tasks and environments. For example, it can reduce the use of thrusters when going downstream and find the route with the least energy consumption when going upstream.

[0111] For example, the robot discovered that there are favorable ocean currents in certain waters, and using a reinforcement learning algorithm, it learned how to use the power of the ocean currents to move forward and reduce energy consumption.

[0112] The specific implementation process of reinforcement learning technology is as follows:

[0113] Reinforcement learning techniques include state space, action space, and reward function, where:

[0114] State space: includes the robot's position, speed, posture, sensor readings, environmental parameters (such as water flow speed and direction), and obstacle information.

[0115] Action space: The operations that the robot can perform, such as adjusting heading, changing speed, changing depth, obstacle avoidance, etc.

[0116] Reward function: Design a reasonable reward mechanism to encourage the robot to complete the task objectives and avoid danger. The reward function includes: positive rewards for approaching the target position, positive rewards for avoiding collisions or successfully avoiding obstacles, and negative rewards for excessive energy consumption or action failure.

[0117] Learning Algorithms: Reinforcement learning algorithms such as Q-learning, Deep Q-Network (DQN), or Policy Gradient can be used to enable the robot to learn in a continuous state and action space.

[0118] Examples of specific application scenarios: Mission objective: The robot needs to autonomously complete the cruise mission in a complex underwater environment, collect environmental data and return to the base. Initial behavior: The robot moves along the preset path, but may encounter unknown obstacles or complex water currents.

[0119] Learning process: When the robot encounters an obstacle, it learns an effective obstacle avoidance strategy by trying different obstacle avoidance actions. In a complex water flow environment, the robot learns how to follow the water flow and reduce energy consumption by adjusting its heading and speed. Optimization results: After multiple learning iterations, the robot can: autonomously plan the optimal path to avoid obstacles and unfavorable water flow areas; optimize energy consumption and extend battery life; in emergencies (such as sensor failure), adjust strategies and continue to complete tasks. Advantages of using reinforcement learning: 1) Strong adaptability: The robot can adjust its strategy in real time according to environmental changes and does not rely on pre-programmed fixed algorithms; 2) Improved autonomy: Reduced dependence on human intervention, able to autonomously complete tasks in unknown or dynamic environments; 3) Performance optimization: Through continuous learning and optimization, the robot can improve the efficiency and safety of task completion.

[0120] The power system is used to provide power for the underwater robot to move and turn, receive instructions from the control system and the navigation system, and perform corresponding movement operations to control the movement of the underwater robot according to the path output by the path planning module. The power system includes a propeller, a steering device, and an intelligent power management system (BMS).

[0121] The thruster is used to provide driving force through a brushless motor; it is driven by a brushless motor with high efficiency and low noise. The thrust is 50N, the speed is 6000rpm, and the power consumption is 100W. The driving force control is performed using the vector eight-thrust layout algorithm. The specific process is as follows: initialize the parameters of all thrusters; obtain the input of the navigation system, including the required movement direction and thrust size; calculate the thrust vector of each thruster based on the navigation input; assign the calculated thrust vector to each thruster; execute the thrust command of the thruster through the control system to achieve omnidirectional movement. Continuously monitor the working status of the thruster to ensure normal operation; dynamically adjust the thrust distribution and optimize motion control based on real-time environmental data and thruster status.

[0122] The steering device is used to control the moving direction of the underwater robot through the servo control system. The servo control system is adopted to ensure precise direction control. The type of servo motor is digital servo, with a control accuracy of ±1 degree and a response speed of 0.1 second.

[0123] The intelligent power management system is used to measure the real-time battery status of the battery, which includes parameters such as voltage, current and temperature; use the left-right symmetrical battery compartment algorithm to ensure that the battery operates within a safe range. The specific process is as follows: initialize the parameters of the battery management system; read battery parameters such as voltage, current and temperature; determine whether the real-time battery status exceeds the battery safety limit; if so, generate a safety alarm and optimize the battery use strategy; optimize battery use efficiency by balancing battery cells; transmit the battery status to the control system to support decision-making.

[0124] The underwater robot dynamic environment perception and navigation device also includes a front flow field information collection system; the front flow field information collection system is used to: collect flow field information of the underwater environment in real time to obtain initial flow field information; the flow field information includes water flow velocity, direction and turbulence characteristics.

[0125] The front flow field information collection system is installed at the front of the underwater robot. The front flow field information collection system includes a micro Doppler flow meter and a flow field sensor. It can sample flow field data at a high frequency, generate a flow field model through a real-time processing module, and feed the data back to the data processing system and control system to assist navigation and obstacle avoidance decisions.

[0126] The data processing system preprocesses the initial flow field information to obtain the flow field information to be detected.

[0127] The flow field information to be detected is updated into the flow field model. If it is detected that the difference between the flow field information to be detected and the set flow field threshold is greater than the set flow field threshold, the output parameters of the navigation system are adjusted.

[0128] The steps of the algorithm of the front flow field information collection system are: initialize a flow field model to store flow field data; read the current flow field data from the front flow field sensor; process and analyze the read flow field data; update the processed flow field data to the flow field model; if a significant change in the flow field is detected, adjust the parameters of the navigation system; store and transmit the flow field data to the control system to support decision-making.

[0129] The underwater robot dynamic environment perception and navigation device also includes a shock absorption system, which reduces the impact of external impact on internal components through elastic materials and energy absorption structure design. The shock absorption system is installed at key parts of the underwater robot and is used to: initialize the damping parameters of the shock absorption system; measure the external force applied to the underwater robot through sensors to obtain the external force; determine whether the external force is greater than the set external force threshold, and if so, calculate the new damping parameter and apply the new damping parameter; if the external force is within the safe range, keep the current damping parameter unchanged; and continuously update sensor data to monitor changes in external force.

[0130] The robot's multi-degree-of-freedom system motion equation is shown below:

[0131] M×{x”(t)}+C×{x’(t)}+K×{x(t)}={F(t)};

[0132] Among them, M is the mass matrix, C is the damping matrix, K is the stiffness matrix, {x(t)} is the displacement vector, {x'(t)} is the velocity vector, {x”(t)} is the acceleration vector, and {F(t)} is the external force vector.

[0133] Since the damping matrix C is usually difficult to determine directly, the assumption of proportional damping (also called Rayleigh damping) is often used to approximate the damping matrix, which is in the form of:

[0134] C=a×M+b×K; wherein a is the damping coefficient related to mass, and b is the damping coefficient related to stiffness.

[0135] To determine a and b, the natural frequency and damping ratio of the system are used. The specific process is as follows:

[0136] (1) Obtain the natural frequency of the system: Calculate the natural frequency of the system, denoted as w1, w2, ..., w n .

[0137] (2) Setting the damping ratio: According to the material properties or experimental data, the damping ratio of the system is set, denoted as zeta.

[0138] (3) Establish the equation system:

[0139] For the i-th mode, the relationship between its damping ratio zeta and the damping coefficients a and b is:

[0140] zeta=[(a / (2×w i ))]+[(b×w i ) / 2]

[0141] Among them, w i is the i-th natural frequency, i=1, 2, ..., n.

[0142] (4) Select two modes and establish equations: Select two natural frequencies w1 and w2, and assume that the corresponding damping ratios are the same, both zeta. Then:

[0143] For mode 1, we obtain equation 1: zeta = [(a / (2×w1))] + [(b×w1) / 2];

[0144] For mode 2, we obtain equation 2: zeta = [(a / (2×w2))] + [(b×w2) / 2].

[0145] (5) By rearranging the above two equations, we can obtain a system of linear equations about a and b, and solving them can give the values ​​of a and b.

[0146] 4. Calculation steps: Assuming that the natural frequencies w1 and w2 are known and the damping ratio zeta is set, the specific solution steps are as follows:

[0147] (1) Rearrange the equation: Multiply the equation by 2 to obtain:

[0148] Equation 1: (a / w1)+b×w1=2×zeta;

[0149] Equation 2: (a / w2)+b×w2=2×zeta.

[0150] (2) Solve the system of equations: Use substitution or elimination to solve the system of equations and find a and b.

[0151] 5. When applied to the underwater robot shock absorption system, in the underwater robot, when the external force is detected to exceed the set threshold, a new damping coefficient needs to be calculated to adjust the damping characteristics of the system. The specific steps are as follows:

[0152] (1) Real-time monitoring of external force: Use sensors to measure the external force acting on the robot, denoted as F.

[0153] (2) Determine whether damping needs to be adjusted: If the external force F is greater than the set external force threshold, a new damping coefficient needs to be calculated.

[0154] (3) Set a new damping ratio: According to actual needs, set a new damping ratio, denoted as zeta_new.

[0155] (4) Recalculate a and b: Use the new damping ratio zeta_new, repeat step 3 above, and solve the equations to obtain new a and b.

[0156] (5) Update the damping matrix:

[0157] Substitute the new a and b into the damping matrix formula: Cnew=a×M+b×K.

[0158] (6) Apply new damping parameters: Apply the new damping matrix Cnew to the shock absorption system to adjust the damping characteristics of the shock absorption system.

[0159] It should be noted that the selected natural frequencies w1 and w2 should represent the main vibration characteristics of the system; the value of the damping ratio zeta should be based on material properties, structural requirements or experimental data; when adjusting the damping parameters, the stability of the system should be ensured and no new vibration problems should be introduced; in practical applications, it may be necessary to quickly calculate the new damping coefficient, so the algorithm should be efficient.

[0160] In another exemplary embodiment of the present application, the underwater robot dynamic environment perception and navigation device also includes a pressure detection system; the pressure detection system includes a pressure processing unit and a plurality of pressure detection points; the pressure detection points use high-precision pressure sensors, which are installed at key parts of the underwater robot shell and interior, and can withstand extreme pressure environments; the pressure detection points are used to measure the real-time pressure of key parts of the underwater robot to ensure the safety of structural design and materials under high-pressure environments; the pressure processing unit is used to: initialize a data structure to store pressure data; read current pressure data from the pressure sensor; for each key part, determine whether the real-time pressure of the key part is greater than the pressure safety limit corresponding to the key part (the safety limit of different pressure detection nodes is different), and if so, generate a safety alarm and record the event; generate a real-time pressure report based on the read real-time pressure data; and transmit the generated pressure report to the control system.

[0161] The sensor data is aggregated and processed through sensor repeaters to form a real-time pressure monitoring report to assist the underwater robot in operational decision-making.

[0162] The communication system of the present application includes acoustic communication, wireless communication and optical fiber communication, ensuring reliable data transmission in different environments. Acoustic communication is used for long-distance underwater communication, wireless communication is used in shallow water areas, and optical fiber communication is used for high data volume transmission.

[0163] Acoustic communication: used for long-distance underwater communication. The main technical parameters include communication frequency, bandwidth and maximum communication distance.

[0164] Wireless communication: used in shallow waters to ensure high-bandwidth data transmission. Including communication frequency, protocol and transmission rate.

[0165] Fiber optic communication: used for high data volume transmission, ensuring low latency and high reliability.

[0166] The specific steps of the data transmission and communication system algorithm are as follows: initialize each communication channel in the communication system, including acoustic communication, wireless communication and optical fiber communication; collect data packets generated by each subsystem of the underwater robot; evaluate the signal quality of the current communication channel to ensure the reliability of data transmission; dynamically optimize the signal path of data transmission based on the signal quality; transmit data packets through the optimal channel to ensure efficient data transmission; record detailed information of each data transmission, including time, channel and data packet content; update communication channels: update and maintain communication channels according to actual usage.

[0167] In another exemplary embodiment of the present application, the underwater robot dynamic environment perception and navigation device also includes a collision buffer component, which is used to: initialize the parameters of the collision buffer component, including energy-absorbing materials and structural designs; use sensors to detect the collision between the underwater robot and obstacles; when a collision occurs, the collision buffer component absorbs the collision energy through energy-absorbing materials and structural designs to reduce the impact on internal components; disperses the absorbed collision energy to various parts of the buffer component to further reduce the impact force; continuously monitors the status of the collision buffer component to ensure its normal operation after the collision; and updates the parameters of the buffer component according to the actual collision situation to cope with possible future collisions.

[0168] The present application adopts a modular design, specifically: initializing the parameters of the modular design, including interface standards and assembly methods; detecting the working status of each module of the underwater robot and recording operating data; identifying and recording faulty modules in the system; maintaining and replacing modules based on the fault detection results; updating module information in the system after maintaining or replacing modules; recording detailed information for each maintenance and replacement operation for subsequent tracking and analysis, and facilitating maintenance and upgrades.

[0169] Through the above modules, the underwater robot of the present application can provide stable and efficient operating performance in unknown environments, and has the advantages of high strength and stability, modular convenience, excellent impact resistance, all-round perception capability, high adaptability and flexibility, and simplified maintenance and upgrades.

[0170] This application has the advantages of high strength and stability, modular convenience, excellent impact resistance, all-round perception, high adaptability and flexibility, and simplified maintenance and upgrades through sensor systems, adaptive algorithms, modular design, and high-strength pressure-resistant structures. By integrating multiple sensors, it can fully perceive the surrounding environment in real time, including sonar, cameras, lidar, DVL, depth sensors, temperature sensors, water quality sensors, and IMU, etc., to improve all-round perception capabilities. Through adaptive algorithms, multi-sensor fusion technology, and innovative collision buffer design, the autonomous navigation and operation capabilities of underwater robots have been significantly improved, overcoming the shortcomings of traditional technologies. Real-time perception and autonomous navigation of complex underwater environments are achieved.

[0171] Based on the same inventive concept, the embodiment of the present application also provides an underwater robot dynamic environment perception and navigation method for realizing the underwater robot dynamic environment perception and navigation device involved above. The implementation scheme for solving the problem provided by the method is similar to the implementation scheme recorded in the above device, so the specific limitations in one or more underwater robot dynamic environment perception and navigation method embodiments provided below can refer to the limitations of the underwater robot dynamic environment perception and navigation device above, and will not be repeated here.

[0172] In an exemplary embodiment, Figure 2 As shown, a method for dynamic environment perception and navigation of an underwater robot is provided, including:

[0173] S1: Detecting the underwater environment in which the underwater robot is located to obtain surrounding environment data; the surrounding environment data includes the speed and position of the underwater robot, the depth of the underwater robot in the underwater environment, the temperature of the underwater environment, the water quality information of the underwater environment, and the movement state and direction of the underwater robot;

[0174] S2: Scan the underwater environment and detect obstacles to obtain long-distance information and short-distance information; the long-distance is a distance falling within a first set threshold range; the short-distance is a distance falling within a second set threshold range;

[0175] S3: Collecting an environmental image of the underwater environment; the long-distance information, the short-distance information and the environmental image constitute underwater obstacle information;

[0176] S4: Preprocessing the surrounding environment data and underwater obstacle information respectively to obtain real-time environment data and obstacle information to be detected;

[0177] S5: The optimal path is calculated by the path planning module according to the real-time environmental data;

[0178] S6: performing obstacle detection according to the environment image to obtain an obstacle detection result; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after receiving the obstacle avoidance decision instruction, the path planning module uses an adaptive algorithm to adjust the optimal path according to the obstacle information to be detected to obtain an adjusted path;

[0179] S7: Control the movement of the underwater robot according to the path output by the path planning module.

[0180] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0181] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A dynamic environment perception and navigation device for an underwater robot, characterized in that: The underwater robot dynamic environment perception and navigation device comprises an underwater robot and a sensor system, a data processing system, a control system, a navigation system and a power system arranged on the underwater robot; The sensor system includes a sonar sensor, a camera, a lidar, a Doppler velocity meter, a depth sensor, a temperature sensor, a water quality sensor and an inertial measurement unit; The sonar sensor is used to scan the underwater environment and detect long-distance obstacles to obtain long-distance information; the long distance is a distance that falls within a first set threshold range; The camera is used to collect environmental images of the underwater environment; The laser radar is used to scan the underwater environment and detect short-range obstacles to obtain short-range information; the short-range is a distance falling within a second set threshold range; the long-range information, the short-range information and the environment image constitute the underwater obstacle information; The Doppler velocity measuring instrument is used to measure the speed and position of the underwater robot; the depth sensor is used to measure the depth of the underwater robot in the underwater environment; the temperature sensor is used to measure the temperature of the underwater environment; the water quality sensor is used to measure the water quality information of the underwater environment; the inertial measurement unit is used to measure the motion state and direction of the underwater robot; the speed and position of the underwater robot, the depth of the underwater robot in the underwater environment, the temperature of the underwater environment, the water quality information of the underwater environment, and the motion state and direction of the underwater robot constitute the surrounding environment data; The data processing system is used to pre-process the surrounding environment data and underwater obstacle information respectively to obtain real-time environment data and obstacle information to be detected; The control system includes a central processing unit for controlling the navigation system and the power system; The navigation system includes a path planning module and a dynamic obstacle avoidance module; the path planning module is used to calculate the optimal path according to the real-time environmental data; the dynamic obstacle avoidance module is used to: detect obstacles according to the environmental image to obtain obstacle detection results; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after receiving the obstacle avoidance decision instruction, the path planning module uses an adaptive algorithm to adjust the optimal path according to the obstacle information to be detected to obtain an adjusted path; The power system is used to provide the underwater robot with power for movement and steering, so as to control the movement of the underwater robot according to the path output by the path planning module.

2. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: The underwater robot dynamic environment perception and navigation device also includes a front flow field information collection system; the front flow field information collection system is used to: collect the flow field information of the underwater environment in real time to obtain initial flow field information; The flow field information includes water flow velocity, direction and turbulence characteristics; The data processing system pre-processes the initial flow field information to obtain the flow field information to be detected; The flow field information to be detected is updated into the flow field model. If it is detected that the difference between the flow field information to be detected and the set flow field threshold is greater than the set flow field threshold, the output parameters of the navigation system are adjusted.

3. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: The underwater robot dynamic environment perception and navigation device also includes a shock absorption system; the shock absorption system is installed at various key positions of the underwater robot and is used to: measure the external force applied to the underwater robot through sensors to obtain the external force; determine whether the external force is greater than a set external force threshold, and if so, calculate a new damping parameter.

4. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: The underwater robot dynamic environment perception and navigation device also includes a pressure detection system; the pressure detection system includes a pressure processing unit and a plurality of pressure detection points; the pressure detection points are installed at key parts of the underwater robot shell and interior; the pressure detection points are used to measure the real-time pressure of key parts of the underwater robot; the pressure processing unit is used to: for each key part, determine whether the real-time pressure of the key part is greater than the pressure safety limit corresponding to the key part, and if so, generate a safety alarm.

5. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: In terms of calculating the optimal path based on the real-time environmental data, the path planning module is used to: use the A* algorithm and the Dijkstra algorithm to calculate the optimal path based on the real-time environmental data.

6. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: The power system includes a propeller and a steering device; the propeller is used to provide driving force through a brushless motor; and the steering device is used to control the moving direction of the underwater robot through a servo control system.

7. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: The power system also includes an intelligent power management system; the intelligent power management system is used to measure the real-time battery status of the battery; the real-time battery status includes voltage, current and temperature; determine whether the real-time battery status exceeds the battery safety limit; if so, generate a safety alarm.

8. The underwater robot dynamic environment perception and navigation device according to claim 1, characterized in that: The underwater robot dynamic environment perception and navigation device also includes a communication system; the sensor system, data processing system, control system, navigation system and power system are connected through the communication system.

9. An underwater robot dynamic environment perception and navigation method based on the underwater robot dynamic environment perception and navigation device according to any one of claims 1 to 8, characterized in that: The underwater robot dynamic environment perception and navigation method comprises: Detecting the underwater environment in which the underwater robot is located to obtain surrounding environment data; the surrounding environment data includes the speed and position of the underwater robot, the depth of the underwater robot in the underwater environment, the temperature of the underwater environment, the water quality information of the underwater environment, and the movement state and direction of the underwater robot; Scanning the underwater environment and detecting obstacles to obtain long-distance information and short-distance information; the long-distance is a distance falling within a first set threshold range; the short-distance is a distance falling within a second set threshold range; Collecting an environmental image of an underwater environment; the long-distance information, the short-distance information and the environmental image constitute underwater obstacle information; Preprocess the surrounding environment data and underwater obstacle information respectively to obtain real-time environment data and obstacle information to be detected; The optimal path is calculated based on real-time environmental data through the path planning module; Obstacle detection is performed according to the environment image to obtain an obstacle detection result; when an obstacle is detected, the control system issues an obstacle avoidance decision instruction, and after the path planning module receives the obstacle avoidance decision instruction, the path planning module uses an adaptive algorithm to adjust the optimal path according to the obstacle information to be detected to obtain an adjusted path; The underwater robot is controlled to move according to the path output by the path planning module.

Images