Self-adaptive dynamic path planning method and system for underwater operation robot

Abstract

The invention discloses a self-adaptive dynamic path planning method and system for an underwater operation robot, relates to the technical field of path planning, and aims to effectively reduce the collision risk and further improve the safety of the underwater operation robot by carefully analyzing and monitoring dynamic and static obstacles and establishing a collision prediction model. According to the method, the safety of equipment and environment is guaranteed, in combination with path consumption energy Xnn, a proper path is conveniently and preliminarily selected, the energy consumption is further reduced, the operation efficiency and cruising ability of the robot are improved, and according to the detour difference Rc, detour of a static obstacle appearing in the path is further analyzed so as to seek the shortest path for detour of the static obstacle; meanwhile, the running state of the dynamic obstacle is mastered in real time according to the obstacle running information, the dangerous dynamic obstacle is further locked, a collision prediction model is established, the environment change can be quickly coped with in the running process of the robot, and a foundation is laid for guaranteeing smooth completion of tasks.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to an adaptive dynamic path planning method and system for an underwater working robot. Background Art

[0002] In the large field of intelligent robots, an important technology in the field of underwater robots is an underwater operation robot, which refers to an intelligent robot that can perform various tasks in an underwater environment, such as seabed exploration, underwater equipment maintenance, and scientific research. The use of robots can simulate human thinking and behavior, and has the ability to autonomously perceive, make decisions, and execute tasks.

[0003] However, this adaptive dynamic path planning method still has some shortcomings. For example, traditional underwater robots may systematically analyze dynamic obstacles and static obstacles, and may ignore the detailed analysis of dynamic obstacles and static obstacles. In particular, the robot's avoidance of dynamic obstacles requires further analysis. In addition, the energy consumption state should be considered for the preliminary robot path. Therefore, the traditional system still needs further improvement and optimization to further improve the safety and efficiency of underwater operation robots. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an adaptive dynamic path planning method and system for an underwater working robot, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for adaptive dynamic path planning of an underwater working robot, comprising the following steps: S1. Preliminarily construct a three-dimensional road condition model for the ecological status at 500 meters underwater, and transmit the three-dimensional road condition model to the simulation platform for simulation operation, determine the target point and several groups of paths, and generate the path consumption energy Xnn based on the several groups of paths, and initially select the optimal path; S2. According to the initially selected optimal path, the obstacle area and the driving area are identified, wherein the obstacle area includes a static obstacle area and a dynamic obstacle area, and the static obstacle area and the dynamic obstacle area are monitored respectively to obtain obstacle size information and obstacle operation information respectively, and a detour difference Rc is generated according to the obstacle size information and the position of the target point, and a detour rule is selected; S3. According to the obstacle operation information, obtain the position coordinates of the dynamic obstacle in real time, determine the distance between the robot and the dynamic obstacle, identify the running direction of the dynamic obstacle, monitor and lock the dangerous dynamic obstacle, establish a collision prediction model based on the dangerous dynamic obstacle, and obtain the time after training. The location of the dangerous dynamic obstacles behind and time The robot's position after , and for the time Location of dangerous dynamic obstacles and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot; S4. According to the estimation in S3, if there is a collision, the robot path is tested using the trained collision prediction model to test the robot's optimization coefficient Yhxs, and the optimization coefficient Yhxs is used as the safety target strategy.

[0006] Preferably, data on the ecological status at 500 meters underwater is collected in advance, and the preprocessed data is summarized to obtain an ecological data set. A three-dimensional road condition model is constructed using Unity simulation software, and the three-dimensional road condition model is transmitted to the simulation platform. According to the three-dimensional road condition model, the target point and several groups of paths for the robot's underwater operation are determined.

[0007] Preferably, several groups of paths are marked with serial numbers, and the power consumption Gh and driving time Xc of the robot in the several groups of paths are monitored and recorded to obtain the path consumption energy Xnn. ; Where n is the number of paths, i=1, 2, 3, ..., n, Gh i It is expressed as the power consumption of the robot in the i-th path, Xc i It is expressed as the driving time of the robot in the i-th path, Xnn i It is represented as the energy consumed by the robot in the i-th path; At the same time, the path energy Xnn consumed by the robot in several groups of paths is recorded and compared, and the minimum value is extracted as the optimal path selected for the first time.

[0008] Preferably, the optimal path selected for the first time is used as the current path option of the robot. According to the three-dimensional road condition model and in combination with the optimal path selected for the first time, the surrounding environment is scanned based on the lidar and the distance information of underwater obstacles is measured to generate three-dimensional point cloud data of the underwater environment. The three-dimensional point cloud data of the underwater environment includes identifying obstacles in the underwater environment, obstacle shapes, positions and distances, and locating and identifying obstacle areas and driving areas. The obstacle areas include static obstacle areas and dynamic obstacle areas, wherein underwater obstacles include but are not limited to reefs, rocks, migratory organisms and seaweed.

[0009] Preferably, according to the locked static obstacle area and dynamic obstacle area, monitoring nodes are deployed in the static obstacle area and the dynamic obstacle area respectively, and obstacle size information is monitored and recorded in the static obstacle area, wherein the obstacle size information includes the basic geometric shape, volume size, path value at the contact cross section and surface characteristics of the static obstacle, wherein the surface characteristics include the smoothness, roughness and multi-angularity of the static obstacle; At the same time, the obstacle operation information is monitored and recorded in the dynamic obstacle area, and the obstacle operation information includes the swimming speed of the dynamic obstacle at each time node during the monitoring period, and the average swimming speed of the dynamic obstacle during the monitoring period is generated. , and simultaneously identify the movement direction of the dynamic obstacle, where the movement direction of the dynamic obstacle includes a positive direction and a reverse direction.

[0010] Preferably, according to the obstacle size information in the static obstacle area and in combination with the position of the target point determined in S1, the point where the robot first contacts the static obstacle and the point where the robot does not contact the static obstacle are locked, and the first detour driving distance Wr1 and the second detour driving distance Wr2 are obtained, and the detour difference Rc is generated, and the detour difference Rc is calculated by the following formula: ; The significance of this formula is to further analyze the driving difference between the two detour paths chosen by the robot when it encounters the current static obstacle; If the detour difference Rc is a positive number, it means that the current robot needs to travel along the path where the detour distance Wr2 of the second detour is located; If the detour difference Rc is zero, it means that the current robot can choose to travel on the path with the first detour distance Wr1 and the path with the second detour distance Wr2. If the detour difference Rc is a negative number, it means that the current robot needs to travel according to the path where the detour distance Wr1 is located.

[0011] Preferably, the positive direction refers to the direction in which the running directions of the dynamic obstacle and the robot have an intersection, and the reverse direction refers to the direction in which the running directions of the dynamic obstacle and the robot do not have an intersection; based on the dynamic obstacles marching in the positive direction, the distance Bj between the robot and the dynamic obstacle is obtained, and according to the trained collision prediction model, the time The location of the dangerous dynamic obstacles behind and time The robot's position after : ; In the formula, It is represented as the position of the dangerous dynamic obstacle at time t, It is expressed as the mean swimming speed of the dynamic obstacle at time t, Represented as from time t to the future time length of time, Represented as the position of the robot at time t, It is expressed as the moving mean of the robot at time t.

[0012] Preferably, the time Location of dangerous dynamic obstacles and time Robot's position Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot: like When , it indicates that the dangerous dynamic obstacle and the robot have a collision overlap point. At this time, the robot will automatically perform avoidance measures and start the planning scheme; like , it means that there is no collision overlap between the dangerous dynamic obstacle and the robot, and the driving state of the robot does not change.

[0013] Preferably, for the collision overlap point between the dangerous dynamic obstacle and the robot, and combined with the distance Bj between the robot and the dangerous dynamic obstacle, an optimization index Yhxs is generated, and a planning scheme is started. The specific planning scheme content is as follows: ; Where m represents the number of dangerous dynamic obstacles, j = 1, 2, 3, ..., m, k attr Expressed as the repulsion coefficient, Bj j It is represented as the distance between the robot and the jth dangerous dynamic obstacle, and Bjv is represented as the safe distance between the robot and the dangerous dynamic obstacle.

[0014] An adaptive dynamic path planning system for underwater operation robots. The construction module constructs a three-dimensional road condition model of the ecological status 500 meters underwater in advance, and transmits the three-dimensional road condition model to the simulation platform for simulation operation to determine the target point and several groups of paths; The preliminary planning module generates the path energy consumption Xnn based on several groups of paths and selects the optimal path for the first time; An information collection module identifies an obstacle area and a driving area according to the initially selected optimal path, wherein the obstacle area includes a static obstacle area and a dynamic obstacle area, and monitors the static obstacle area and the dynamic obstacle area respectively to obtain obstacle size information and obstacle operation information respectively; The static analysis module generates the detour difference Rc and selects the detour rule based on the obstacle size information and the location of the target point; The dynamic analysis module obtains the position coordinates of dynamic obstacles in real time according to the obstacle operation information, determines the distance between the robot and the dynamic obstacle, identifies the running direction of the dynamic obstacle, monitors and locks the dangerous dynamic obstacle, establishes a collision prediction model based on the dangerous dynamic obstacle, and obtains the time after training. The location of the dangerous dynamic obstacles behind and time The robot's position after , and for the time The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot; The planning management module, based on the estimation in S3, uses the trained collision prediction model to test the robot path if it exists, to test the optimization coefficient Yhxs of the robot, and uses the optimization coefficient Yhxs as the safety target strategy.

[0015] The present invention provides an adaptive dynamic path planning method and system for an underwater working robot, which has the following beneficial effects: (1) Through detailed analysis and monitoring of dynamic and static obstacles and the establishment of a collision prediction model, the risk of collision can be effectively reduced, the safety of underwater working robots can be further improved, and the safety of equipment and the environment can be guaranteed. Combined with the path energy consumption Xnn, it is convenient to initially select a suitable path, further reduce energy consumption, and improve the operation efficiency and endurance of the robot. According to the detour difference Rc, the detour of static obstacles appearing in the path is further analyzed to find the shortest path to bypass the static obstacle. At the same time, according to the obstacle operation information, the operation status of dynamic obstacles is grasped in real time, and dangerous dynamic obstacles are further locked. A collision prediction model is established, which can quickly respond to environmental changes during the operation of the robot and lay the foundation for the smooth completion of the task. Optimizing path planning and safety strategies and testing the robot's optimization coefficient Yhxs can further reduce the probability of collision between the robot and obstacles, reduce unnecessary pauses and adjustments, improve the efficiency and success rate of task completion, and further realize the safe and efficient operation of underwater working robots.

[0016] (2) By collecting and preprocessing the underwater ecological status data in advance, constructing an ecological data set, and using Unity simulation software to build a three-dimensional road condition model, data-driven path planning is further realized. This method is based on real ecological data and simulated three-dimensional road condition models, and can more accurately determine the target point and optimal path of the robot's underwater operation. At the same time, by analyzing the detour difference Rc, it is possible to accurately analyze the driving differences of the robot in real time when encountering static obstacles. This real-time path adjustment enables the robot to always choose the optimal driving path to ensure the smooth completion of underwater operation tasks.

[0017] (3) By predicting whether the dangerous dynamic obstacle and the robot will collide and overlap, the robot can take timely and automatic avoidance measures and start planning. This intelligent avoidance capability can effectively reduce the risk of collision and improve the safety of underwater operation robots. By comparing the position of the dangerous dynamic obstacle and the position of the robot in real time, the collision risk can be judged in time and corresponding actions can be taken. This real-time feedback mechanism enables the robot to respond quickly when encountering emergencies and ensure the smooth progress of underwater operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of the steps of a method for adaptive dynamic path planning of an underwater working robot according to the present invention; Figure 2 The present invention is a schematic diagram of the process flow of an underwater working robot adaptive dynamic path planning system. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0020] Example 1 See also Figure 1 The present invention provides an adaptive dynamic path planning method for an underwater working robot, comprising the following steps: S1. Preliminarily construct a three-dimensional road condition model for the ecological status at 500 meters underwater, and transmit the three-dimensional road condition model to the simulation platform for simulation operation, determine the target point and several groups of paths, and generate the path consumption energy Xnn based on the several groups of paths, and initially select the optimal path; S2. According to the initially selected optimal path, the obstacle area and the driving area are identified, wherein the obstacle area includes a static obstacle area and a dynamic obstacle area, and the static obstacle area and the dynamic obstacle area are monitored respectively to obtain obstacle size information and obstacle operation information respectively, and a detour difference Rc is generated according to the obstacle size information and the position of the target point, and a detour rule is selected; S3. According to the obstacle operation information, obtain the position coordinates of the dynamic obstacle in real time, determine the distance between the robot and the dynamic obstacle, identify the running direction of the dynamic obstacle, monitor and lock the dangerous dynamic obstacle, establish a collision prediction model based on the dangerous dynamic obstacle, and obtain the time after training. The location of the dangerous dynamic obstacles behind and time The robot's position after , and for the time The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot; S4. According to the estimation in S3, if there is a collision, the robot path is tested using the trained collision prediction model to test the robot's optimization coefficient Yhxs, and the optimization coefficient Yhxs is used as the safety target strategy.

[0021] In this embodiment, by pre-establishing a three-dimensional road condition model of the ecological state and performing a simulation operation, the optimal path and energy consumption state can be quickly determined, thereby improving the accuracy and efficiency of path planning; by identifying dynamic and static obstacle areas and monitoring the size, operation information and motion state of the obstacles, the robot can more accurately avoid collision risks and further improve operational safety. Based on the location of dangerous dynamic obstacles and the location of the robot, collision prediction can be performed to optimize the robot's path and formulate a safety target strategy, thereby improving the robot's operational safety, and timely responding to environmental changes and adaptively adjusting the path to ensure the smooth completion of the task. In short, this method combines advanced simulation technology, obstacle identification, collision prediction and path optimization strategies, which can effectively improve the operating efficiency, safety and reliability of underwater operation robots.

[0022] Example 2 Please refer to Figure 1Specifically: collect data on the ecological status 500 meters underwater in advance, summarize the preprocessed data, obtain the ecological data set, and use Unity simulation software to build a three-dimensional road condition model. The three-dimensional road condition model is transmitted to the simulation platform, and the effectiveness and feasibility of path planning can be verified through the simulation platform. This method can verify and adjust the path planning before actual operation to ensure the safety and efficient execution of the robot in actual underwater operations; according to the three-dimensional road condition model, determine the target point of the robot's underwater operation and several groups of paths.

[0023] Several groups of paths are marked with serial numbers, and the power consumption Gh and driving time Xc of the robot in several groups of paths are monitored and recorded to obtain the path consumption energy Xnn. ; Where n is the number of paths, i=1, 2, 3, ..., n, Gh i It is expressed as the power consumption of the robot in the i-th path, Xc i It is expressed as the driving time of the robot in the i-th path, Xnn i It is represented as the energy consumed by the robot in the i-th path; The power consumption Gh mentioned above can be directly measured by a power sensor to measure the power consumed by the robot; The driving time Xc can be calculated by using a timer to record the time when the robot starts and ends the task; At the same time, the path energy consumption Xnn of the robot in several groups of paths is recorded and compared, and the minimum value is extracted as the optimal path selected for the first time. This method can optimize the path selection of the robot, reduce energy consumption, and improve work efficiency and energy saving; In this embodiment, by collecting and preprocessing data and constructing an ecological data set, an accurate description of the underwater ecological state can be further obtained, which improves the accuracy and reliability of path planning; a three-dimensional road condition model is constructed using simulation software, and path planning and optimization are performed in a simulation platform, which saves time and cost of field tests and improves work efficiency. By recording the power consumption and driving time of the robot in different paths, obtaining the path energy consumption data, and comparing the energy consumption of each path, the optimal path can be selected, which optimizes the path planning and driving efficiency of the robot.

[0024] Example 3 Please refer to Figure 1Specifically: the optimal path selected for the first time is used as the current path option of the robot, and according to the three-dimensional road condition model and in combination with the optimal path selected for the first time, the surrounding environment is scanned based on the lidar and the distance information of underwater obstacles is measured to generate three-dimensional point cloud data of the underwater environment, wherein the three-dimensional point cloud data of the underwater environment includes identifying obstacles in the underwater environment, obstacle shapes, positions and distances, and locating and identifying obstacle areas and driving areas, wherein the obstacle areas include static obstacle areas and dynamic obstacle areas, wherein underwater obstacles include but are not limited to reefs, rocks, migratory organisms and seaweed.

[0025] According to the locked static obstacle area and dynamic obstacle area, monitoring nodes are deployed in the static obstacle area and the dynamic obstacle area respectively, and obstacle size information is monitored and recorded in the static obstacle area. The obstacle size information includes the basic geometric shape, volume size, path value at the contact cross section and surface characteristics of the static obstacle. The surface characteristics include the smoothness, roughness and multi-angle conditions of the static obstacle, which may affect the damage of the robot after contact with it; At the same time, the obstacle operation information is monitored and recorded in the dynamic obstacle area, and the obstacle operation information includes the swimming speed of the dynamic obstacle at each time node during the monitoring period, and the average swimming speed of the dynamic obstacle during the monitoring period is generated. , and simultaneously identify the movement direction of the dynamic obstacle, where the movement direction of the dynamic obstacle includes a positive direction and a reverse direction.

[0026] In this embodiment, a laser radar is used to scan the underwater environment and measure the distance information of obstacles to generate three-dimensional point cloud data of the underwater environment. It can accurately identify static obstacles and dynamic obstacles in the underwater environment, including reefs, rocks, swimming organisms and seaweed, and simultaneously locate and identify obstacle areas and driving areas; monitoring nodes are deployed in the static obstacle area to monitor and record obstacle size information, including the geometric shape, volume, path value at the contact cross section and surface characteristics of the static obstacles, so as to fully understand the characteristics of the obstacles and factors that may affect the robot; the obstacle operation information is monitored and recorded in the dynamic obstacle area, including the swimming speed of the dynamic obstacles during the monitoring time period, and the swimming speed average is generated. , and identify the movement direction of dynamic obstacles, such as the forward and reverse directions, so as to timely understand the running status of dynamic obstacles and provide data support for robot path planning and obstacle avoidance; through comprehensive monitoring of the underwater environment and obstacle information recording, it can accurately assess the contact risk and damage between the robot and the obstacle, thereby optimizing the path planning and improving the safety and efficiency of the robot operation. In short, this underwater operation robot path planning method based on LiDAR technology can effectively improve the safety, accuracy and efficiency of the operation, and provide reliable technical support and guarantee for the smooth completion of underwater operation tasks.

[0027] Example 4 Please refer to Figure 1 Specifically: according to the obstacle size information in the static obstacle area and in combination with the position of the target point determined in S1, the robot's initial contact with the static obstacle and the robot's final contact with the static obstacle are locked, and the first detour driving distance Wr1 and the second detour driving distance Wr2 are obtained, and the detour difference Rc is generated. The detour difference Rc is obtained by the following formula: ; The significance of this formula is to further analyze the driving difference between the two detour paths chosen by the robot when it encounters the current static obstacle; The above-mentioned detour No. 1 driving distance Wr1 and detour No. 2 driving distance Wr2 are obtained by taking a camera or a camera to capture the circumference of the contact point between the obstacle and the robot, and combining the point where the robot first contacts the static obstacle and the point where it does not contact the static obstacle. If the detour difference Rc is a positive number, it means that the current robot needs to travel along the path where the detour distance Wr2 of the second detour is located; If the detour difference Rc is zero, it means that the current robot can choose to travel on the path with the first detour distance Wr1 and the path with the second detour distance Wr2. If the detour difference Rc is a negative number, it means that the current robot needs to travel according to the path where the detour distance Wr1 is located.

[0028] In this embodiment, according to the obstacle size information and the target point position in the static obstacle area, combined with the points where the robot first contacts and does not contact the static obstacle, the path options of the robot when encountering the static obstacle can be analyzed, and the No. 1 detour driving distance Wr1 and the No. 2 detour driving distance Wr2 are analyzed and generated, and then the detour difference Rc is calculated. The detour difference Rc is calculated and analyzed by a formula. If Rc is a positive number, it means that the robot should choose the No. 2 detour driving path; if Rc is zero, it means that the robot can choose between the No. 1 and No. 2 detour driving paths; if Rc is a negative number, it means that the robot should choose the No. 1 detour driving path. This detour difference Rc analysis provides a specific basis and guidance for path selection; by analyzing the detour difference Rc, the robot can select the most optimized driving path according to the actual situation, avoid unnecessary driving distance output, and further improve driving efficiency and save energy; at the same time, selecting a suitable driving path can indirectly reduce the contact opportunities and risks between the robot and the static obstacle, reduce the possibility of collision and damage, and thus improve the safety and reliability of the operation. In short, this path planning method based on detour difference Rc analysis can help the robot choose the most suitable and optimized driving path when facing static obstacles, thereby improving work efficiency and safety.

[0029] Example 5 Please refer to Figure 1 Specifically: the positive direction refers to the direction where the running direction of the dynamic obstacle and the running direction of the robot intersect, and the reverse direction refers to the direction where the running direction of the dynamic obstacle and the running direction of the robot do not intersect; based on the dynamic obstacles marching in the positive direction, the distance Bj between the robot and the dynamic obstacle is obtained, and according to the trained collision prediction model, the time The location of the dangerous dynamic obstacles behind and time The robot's position after : ; In the formula, It is represented as the position of the dangerous dynamic obstacle at time t, It is expressed as the mean swimming speed of the dynamic obstacle at time t, Represented as from time t to the future time length of time, It is represented as the position of the robot at time t, It is represented as the moving mean of the robot at time t.

[0030] The moment The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot: like When , it indicates that the dangerous dynamic obstacle and the robot have a collision overlap point. At this time, the robot will automatically perform avoidance measures and start the planning scheme; like , it means that there is no collision overlap between the dangerous dynamic obstacle and the robot, and the driving state of the robot does not change.

[0031] In this embodiment, the possibility of collision is analyzed according to the running direction of the dynamic obstacle and the running direction of the robot; through the distance Bj between the robot and the dynamic obstacle, and using the collision prediction model to perform real-time collision prediction, when the dangerous dynamic obstacle and the robot have a collision overlap point, the robot will automatically execute avoidance measures and start the planning scheme. This automatic avoidance measure can ensure that there is a safe distance between the robot and the dynamic obstacle, further reducing the possibility of collision and damage; when no collision overlap point is found, the driving state of the robot does not change, and the normal driving trajectory and speed are maintained. This driving state control can avoid unnecessary avoidance actions and improve the driving efficiency of the robot. In short, the collision prediction and avoidance measures based on the running direction of the dynamic obstacle and the running direction of the robot can predict the collision risk in real time, automatically execute avoidance measures, ensure the safe distance between the robot and the dynamic obstacle, and improve the safety and efficiency of underwater operations.

[0032] Example 6 Please refer to Figure 1 Specifically: for the collision overlap point between the dangerous dynamic obstacle and the robot, and combined with the distance Bj between the robot and the dangerous dynamic obstacle, generate the optimization index Yhxs, and start the planning scheme. The specific planning scheme content is as follows: ; Where m represents the number of dangerous dynamic obstacles, j = 1, 2, 3, ..., m, k attr Expressed as the repulsion coefficient, Bj j It is represented as the distance between the robot and the jth dangerous dynamic obstacle, and Bjv is represented as the safe distance between the robot and the dangerous dynamic obstacle. The safe distance Bjv between the robot and the dangerous dynamic obstacle is a preset safe distance and a fixed value.

[0033] Among them, for the repulsion coefficient k attrThere are several ways to obtain it: experimental measurement, which can be used to measure the interaction force between the underwater robot and the obstacle. For example, a spring dynamometer and other equipment can be used to place the underwater robot in a controlled experimental environment and measure the repulsion between the robot and the obstacle. The experiment can be carried out in a pool, a sink or other underwater simulation environment; numerical simulation, which can be performed using a computer program to simulate the interaction between the underwater robot and the obstacle; finally, scene simulation can be used. Virtual reality technology can be used to perform scene simulation tests to simulate the movement and interaction of underwater robots in different environments. The repulsion coefficient between the underwater robot and the obstacle can be inferred through the simulation test results.

[0034] In this embodiment, by considering the distance between the robot and the dangerous dynamic obstacle and the safe distance, combined with the repulsion coefficient k attr , generate the optimization index Yhxs, this index can comprehensively consider the collision risk and safety distance, so that the robot can choose the safest path and action when encountering dangerous dynamic obstacles, greatly improving the safety of operation; according to the generated optimization index Yhxs, start the intelligent planning scheme, this planning scheme is flexibly adjusted according to the distance and safety distance between the robot and each obstacle, so that the robot can autonomously select the optimal path in a complex environment, further improving the operation efficiency; the planning scheme is real-time and can be dynamically adjusted according to the real-time position and distance information of the robot and the dangerous dynamic obstacles. This real-time response capability enables the robot to quickly adapt to environmental changes, avoid potential collision risks in time, and ensure the continuity and stability of operations. At the same time, the activation of the intelligent planning scheme is based on the collision overlap point between the dangerous dynamic obstacle and the robot, which effectively alleviates the stability and efficiency of the overall system and reduces the system paralysis or accidents caused by system analysis of multi-source data.

[0035] Example 7 Please refer to Figure 1 and Figure 2 ,Specifically: An adaptive dynamic path planning system for underwater operation robots, The construction module constructs a three-dimensional road condition model of the ecological status 500 meters underwater in advance, and transmits the three-dimensional road condition model to the simulation platform for simulation operation to determine the target point and several groups of paths; The preliminary planning module generates the path energy consumption Xnn based on several groups of paths and selects the optimal path for the first time; An information collection module identifies an obstacle area and a driving area according to the initially selected optimal path, wherein the obstacle area includes a static obstacle area and a dynamic obstacle area, and monitors the static obstacle area and the dynamic obstacle area respectively to obtain obstacle size information and obstacle operation information respectively; The static analysis module generates the detour difference Rc and selects the detour rule based on the obstacle size information and the location of the target point; The dynamic analysis module obtains the position coordinates of dynamic obstacles in real time according to the obstacle operation information, determines the distance between the robot and the dynamic obstacle, identifies the running direction of the dynamic obstacle, monitors and locks the dangerous dynamic obstacle, establishes a collision prediction model based on the dangerous dynamic obstacle, and obtains the time after training. The location of the dangerous dynamic obstacles behind and time The robot's position after , and for the time The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot; The planning management module, based on the estimation in S3, uses the trained collision prediction model to test the robot path if it exists, to test the optimization coefficient Yhxs of the robot, and uses the optimization coefficient Yhxs as the safety target strategy.

[0036] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An adaptive dynamic path planning method for an underwater working robot, characterized in that: The following steps are involved: S1. Preliminarily construct a three-dimensional road condition model for the ecological status at 500 meters underwater, and transmit the three-dimensional road condition model to the simulation platform for simulation operation, determine the target point and several groups of paths, and generate the path consumption energy Xnn based on the several groups of paths, and initially select the optimal path; S2. According to the initially selected optimal path, the obstacle area and the driving area are identified, wherein the obstacle area includes a static obstacle area and a dynamic obstacle area, and the static obstacle area and the dynamic obstacle area are monitored respectively to obtain obstacle size information and obstacle operation information respectively, and a detour difference Rc is generated according to the obstacle size information and the position of the target point, and a detour rule is selected; S3. According to the obstacle operation information, obtain the position coordinates of the dynamic obstacle in real time, determine the distance between the robot and the dynamic obstacle, identify the running direction of the dynamic obstacle, monitor and lock the dangerous dynamic obstacle, establish a collision prediction model based on the dangerous dynamic obstacle, and obtain the time after training. The location of the dangerous dynamic obstacles behind and time The robot's position after , and for the time The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot; S4. According to the estimation in S3, if there is a collision, the robot path is tested using the trained collision prediction model to test the robot's optimization coefficient Yhxs, and the optimization coefficient Yhxs is used as the safety target strategy.

2. The method for adaptive dynamic path planning of an underwater working robot according to claim 1, characterized in that: Data on the ecological status at 500 meters underwater is collected in advance, and the preprocessed data is summarized to obtain the ecological data set. The Unity simulation software is used to construct a three-dimensional road condition model, which is then transmitted to the simulation platform. Based on the three-dimensional road condition model, the target points and several groups of paths for the robot's underwater operations are determined.

3. The method for adaptive dynamic path planning of an underwater working robot according to claim 1, characterized in that: Several groups of paths are marked with serial numbers, and the power consumption Gh and driving time Xc of the robot in several groups of paths are monitored and recorded to obtain the path consumption energy Xnn. ； Where n is the number of paths, i=1, 2, 3, ..., n, Gh i It is expressed as the power consumption of the robot in the i-th path, Xc i It is expressed as the driving time of the robot in the i-th path, Xnn i It is represented as the energy consumed by the robot in the i-th path; At the same time, the path energy Xnn consumed by the robot in several groups of paths is recorded and compared, and the minimum value is extracted as the optimal path selected for the first time.

4. The method for adaptive dynamic path planning of an underwater working robot according to claim 3, characterized in that: The optimal path selected for the first time is used as the current path option of the robot. According to the three-dimensional road condition model and in combination with the optimal path selected for the first time, the surrounding environment is scanned based on the lidar and the distance information of underwater obstacles is measured to generate three-dimensional point cloud data of the underwater environment. The three-dimensional point cloud data of the underwater environment includes identifying obstacles in the underwater environment, obstacle shapes, positions and distances, and locating and identifying obstacle areas and driving areas. The obstacle areas include static obstacle areas and dynamic obstacle areas, wherein underwater obstacles include but are not limited to reefs, rocks, migratory organisms and seaweed.

5. The method for adaptive dynamic path planning of an underwater working robot according to claim 4, characterized in that: According to the locked static obstacle area and dynamic obstacle area, monitoring nodes are deployed in the static obstacle area and the dynamic obstacle area respectively, and obstacle size information is monitored and recorded in the static obstacle area, wherein the obstacle size information includes the basic geometric shape, volume size, path value at the contact cross section and surface characteristics of the static obstacle, wherein the surface characteristics include the smoothness, roughness and multi-angularity of the static obstacle; At the same time, the obstacle operation information is monitored and recorded in the dynamic obstacle area, and the obstacle operation information includes the swimming speed of the dynamic obstacle at each time node during the monitoring period, and the average swimming speed of the dynamic obstacle during the monitoring period is generated. , and simultaneously identify the movement direction of the dynamic obstacle, where the movement direction of the dynamic obstacle includes a positive direction and a reverse direction.

6. The method for adaptive dynamic path planning of an underwater working robot according to claim 5, characterized in that: According to the obstacle size information in the static obstacle area and the position of the target point determined in S1, the robot's initial contact with the static obstacle and the point where the robot did not contact the static obstacle are locked, and the first detour driving distance Wr1 and the second detour driving distance Wr2 are obtained, and the detour difference Rc is generated. The detour difference Rc is calculated by the following formula: ； The significance of this formula is to further analyze the driving difference between the two detour paths chosen by the robot when it encounters the current static obstacle; If the detour difference Rc is a positive number, it means that the current robot needs to travel along the path where the detour distance Wr2 of the second detour is located; If the detour difference Rc is zero, it means that the current robot can choose to travel on the path with the first detour distance Wr1 and the path with the second detour distance Wr2. If the detour difference Rc is a negative number, it means that the current robot needs to travel according to the path where the detour distance Wr1 is located.

7. The method for adaptive dynamic path planning of an underwater working robot according to claim 5, characterized in that: The positive direction refers to the direction in which the running direction of the dynamic obstacle and the running direction of the robot intersect, and the reverse direction refers to the direction in which the running direction of the dynamic obstacle and the running direction of the robot do not intersect; based on the dynamic obstacles marching in the positive direction, the distance Bj between the robot and the dynamic obstacle is obtained, and according to the trained collision prediction model, the time The location of the dangerous dynamic obstacles behind and time The robot's position after : ； In the formula, It is represented as the position of the dangerous dynamic obstacle at time t, It is expressed as the mean swimming speed of the dynamic obstacle at time t, Represented as from time t to the future time length of time, Represented as the position of the robot at time t, It is expressed as the moving mean of the robot at time t.

8. The method for adaptive dynamic path planning of an underwater working robot according to claim 7, characterized in that: The moment The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot: like When , it indicates that the dangerous dynamic obstacle and the robot have a collision overlap point. At this time, the robot will automatically perform avoidance measures and start the planning scheme; like , it means that there is no collision overlap between the dangerous dynamic obstacle and the robot, and the driving state of the robot does not change.

9. The method for adaptive dynamic path planning of an underwater working robot according to claim 8, characterized in that: For the collision overlap point between the dangerous dynamic obstacle and the robot, and combined with the distance Bj between the robot and the dangerous dynamic obstacle, the optimization index Yhxs is generated and the planning scheme is started. The specific planning scheme content is as follows: ； Where m represents the number of dangerous dynamic obstacles, j = 1, 2, 3, ..., m, k attr Expressed as the repulsion coefficient, Bj j It is represented as the distance between the robot and the jth dangerous dynamic obstacle, and Bjv is represented as the safe distance between the robot and the dangerous dynamic obstacle.

10. An adaptive dynamic path planning system for an underwater working robot, comprising an adaptive dynamic path planning method for an underwater working robot according to any one of claims 1 to 9, characterized in that: The construction module constructs a three-dimensional road condition model of the ecological status 500 meters underwater in advance, and transmits the three-dimensional road condition model to the simulation platform for simulation operation to determine the target point and several groups of paths; The preliminary planning module generates the path energy consumption Xnn based on several groups of paths and selects the optimal path for the first time; An information collection module identifies an obstacle area and a driving area according to the initially selected optimal path, wherein the obstacle area includes a static obstacle area and a dynamic obstacle area, and monitors the static obstacle area and the dynamic obstacle area respectively to obtain obstacle size information and obstacle operation information respectively; The static analysis module generates the detour difference Rc and selects the detour rule based on the obstacle size information and the location of the target point; The dynamic analysis module obtains the position coordinates of dynamic obstacles in real time according to the obstacle operation information, determines the distance between the robot and the dynamic obstacle, identifies the running direction of the dynamic obstacle, monitors and locks the dangerous dynamic obstacle, establishes a collision prediction model based on the dangerous dynamic obstacle, and obtains the time after training. The location of the dangerous dynamic obstacles behind and time The robot's position after , and for the time The location of the dangerous dynamic obstacles behind and time The robot's position after Compare and estimate whether there is a collision overlap between the dangerous dynamic obstacle and the robot; The planning management module, based on the estimation in S3, uses the trained collision prediction model to test the robot path if it exists, to test the optimization coefficient Yhxs of the robot, and uses the optimization coefficient Yhxs as the safety target strategy.

Images