A Search and Rescue ROV Control System and Method with Path Planning Function

Abstract

The present invention discloses a search and rescue ROV control system and method with path planning function, including an underwater environment grid map model establishment module for constructing an underwater environment grid map model through the working environment of the underwater robot; a path planning module for planning a good path by using an improved genetic algorithm. The process of the improved genetic algorithm for path planning is as follows: initializing the population by using the A* algorithm, designing a path fitness function, and performing crossover, mutation, and selection operations on the path; a path optimization module for optimizing the good path by using a Bezier optimization operator to obtain an optimal path; and a control module for controlling the underwater robot to perform search and rescue through the optimal path. The present invention ensures the safe and efficient movement of the robot by introducing the A* algorithm to optimize the initial population and designing a fitness function with a safety distance; and utilizes the smoothness of the Bezier curve to make the path smoother and more coherent, reducing the energy loss during the movement of the robot and improving the search and rescue efficiency.

Description

Technical Field

[0001] The present invention relates to the field of rescue, and particularly to a search and rescue ROV control system and method with path planning function. Background Art

[0002] ROVs have long been applied to rescue work. They can work under various water depth conditions from the sea surface to the seabed. However, in the actual rescue process, the search work is the premise of rescue. Usually, autonomous underwater vehicles (AUVs) are used to conduct large-scale scans in the accident area. When the lost object is found, the AUV returns, and the ROV is replaced to carry out the salvage operation on the lost object. This rescue method that divides the search and rescue into two parts obviously cannot meet the requirements of the rapid rescue mission.

[0003] A path planning method for a mobile robot based on an improved genetic algorithm proposed by Zhao Jing et al. of Nanjing University of Posts and Telecommunications discloses a path planning method for a mobile robot based on an improved genetic algorithm. The steps include: using a grid map to model the environment of the robot's moving space; setting algorithm parameters and initializing the population; constructing a fitness function using the path length function and the smoothness function; introducing an elite retention strategy, that is, when performing roulette wheel selection, the optimal individual is retained to the next generation and the crossover and mutation operations are continued; dynamically adjusting the population using an adaptive crossover rate and mutation rate; determining whether the number of evolutions reaches the maximum. If so, the optimal solution is output. If not, the above steps are repeated. This invention not only enhances the ability of the mobile robot to find the optimal solution but also improves its convergence speed and reduces the number of turns when applied to the path planning of a mobile robot. However, it has many deficiencies for search and rescue underwater robots, especially search and rescue ROV robots. Due to the limitation of the umbilical cable of the ROV robot itself, the obstacle avoidance ability of the robot is required to be higher. The above technical solution has no excessive requirements in this regard, and it is very likely to cause the entanglement of the umbilical cable when applied to the ROV, which is not conducive to the search and rescue work. Summary of the Invention

[0004] Object of the Invention: The object of the present invention is to provide a search and rescue ROV control system that can stay away from obstacles, has a better initial path, and has a path with better quality and smoothness. Another object of the present invention is to provide a search and rescue ROV control method.

[0005] Technical solution: The search and rescue ROV control system of the present invention includes an underwater environment grid map model establishment module for constructing an underwater environment grid map model through the working environment of the underwater robot; a good path is obtained by planning the path using an improved genetic algorithm. The process of planning the path using the improved genetic algorithm is as follows: initializing the population using the A* algorithm, designing a path fitness function, and performing crossover, mutation, and selection operations on the path; a path optimization module for optimizing the good path using a Bezier optimization operator to obtain an optimal path; a control module for controlling the underwater robot to perform search and rescue through the optimal path.

[0006] Further, in the underwater environment grid map model establishment module,

[0007] In the underwater environment grid map model, black grids represent obstacles, and white grids represent the free space of the underwater robot. The grid is marked using the serial number marking method, and the numbers are incremented by 1 in sequence according to the principle from left to right and from bottom to top; in the underwater environment grid map model, S, 2, 3,..., E are grid nodes, and each grid node is a path point; among them, S is the starting position of the underwater robot, and E is the ending position of the underwater robot; the relationship between the given grid serial number T(i, j) and the rectangular coordinate point P(i, j)[X(i, j), Y(i, j)] is as follows:

[0008]

[0009] Among them, mod represents the behavior of the modulo algorithm, and floor represents the behavior of integer division.

[0010] Further, in the path planning module,

[0011] Initializing the population using the A* algorithm: randomly selecting the starting coordinate and the ending coordinate, and adding the starting coordinate and the ending coordinate to the chromosome encoding; adding an obstacle-free random point as the target point in the underwater environment grid map model; using the A* algorithm to establish the paths from the starting coordinate point to the target point and from the target point to the ending coordinate point respectively, and avoiding obstacle points and peripheral areas during the path establishment process; connecting the starting coordinate point and the ending coordinate point, and deleting the path nodes that are exactly the same, and finally generating a path set;

[0012] Designing a path fitness function: sampling each path point in the path set, calculating the distance from each path point to the nearest obstacle point, when the distance is less than the value P, the fitness of the path increases; when the distance is greater than or equal to the value P, the fitness of the path decreases, where P is greater than or equal to half of the grid side length of the underwater environment grid map model; selecting the path with the minimum fitness as the good path:

[0013] Fitness(i) = ω1Length(p) + ω2Num

[0014] Among them, Fitness(i) is the fitness function, Length(p) is the path length, and Num is the number of path points on the path whose distance from the obstacle point is less than P; ω1 and ω2 are the proportionality coefficients of the path length and the number of path points close to the obstacle. When ω1 > ω2, it means that ensuring a shorter path length is more prioritized. When ω1 < ω2, it means that staying away from the obstacle to ensure operation safety is more prioritized;

[0015] Perform crossover and mutation operations on the path: The individual is the path, and one individual is a single path. Perform crossover and mutation operations to generate more offspring individuals;

[0016] The crossover probability is represented by P c In path planning, the crossover operation means swapping the two parent individuals found at the intersection point. After swapping, keep the shorter path and discard the longer path; After the crossover operation, the fitness of the offspring individual is lower than that of the parent individual;

[0017] The mutation probability is represented by P m In path planning, the mutation operation means flipping the found parent individual with probability P m and then combining with the crossover operation to obtain an offspring individual with a lower fitness;

[0018] The adaptive model of the crossover probability and the mutation probability is:

[0019]

[0020]

[0021] Among them, P c and P m are the crossover probability and the mutation probability respectively; are the maximum crossover probability and the minimum crossover probability of each generation respectively, are the maximum mutation probability and the minimum mutation probability of each generation respectively; f is the fitness function value of each generation, f avg is the average value of the fitness function, f max is the maximum value of the average value of the fitness function, f min is the minimum value of the average value of the fitness function;

[0022] When the fitness of the individual is lower than the average fitness, the crossover and mutation probabilities of the individual remain at the maximum value When the fitness of the individual is higher than the average fitness, the greater the fitness, the lower the crossover and mutation probabilities;

[0023] Perform a path selection operation: Select the most elite individual with the lowest fitness in the parent population as the good path, and retain the good path for comparison with the most elite individual in the offspring population. If the fitness of the offspring most elite individual is less than the fitness of the good path, update the offspring most elite individual to the good path; otherwise, still retain the parent most elite individual as the good path.

[0024] Further, in the path optimization module,

[0025] Assume that there is a Bezier curve containing m control points P0, P1, P2, …, P m The mathematical expression of this curve is as follows:

[0026]

[0027] where m ∈ N+, t is the normalized time variable, P i =(x i , y i ) T is the coordinate vector of the i-th control point, and x i and y i are the components of the X and Y coordinates respectively, is the Bernstein polynomial, which is the basic function of the Bezier curve expression, and its expansion is: Use the mathematical expression of the above Bezier curve to optimize the good path to obtain the optimal path.

[0028] The search and rescue ROV control method of the present invention includes:

[0029] (1) Construct an underwater environment grid map model through the working environment of the underwater robot;

[0030] (2) Obtain a good path by planning the path using an improved genetic algorithm. The process of planning the path using the improved genetic algorithm is: Initialize the population using the A* algorithm, design a path fitness function, and perform crossover, mutation, and selection operations on the path;

[0031] (3) Optimize the good path using a Bezier optimization operator to obtain the optimal path;

[0032] (4) Control the underwater robot to perform search and rescue through the optimal path.

[0033] Further, in step (1),

[0034] In the underwater environment grid map model, black grids represent obstacles, and white grids represent the free space of the underwater robot. The grid labeling is carried out using the serial number labeling method, and the labels are incremented by 1 in sequence according to the principle from left to right and from bottom to top. In the underwater environment grid map model, S, 2, 3, …, E are grid nodes, and each grid node is a path point; among them, S is the starting position of the underwater robot, and E is the ending position of the underwater robot; the relationship between the given grid serial number T(i, j) and the rectangular coordinate point P(i, j)[X(i, j), Y(i, j)] is as follows:

[0035]

[0036] Among them, mod represents the behavior of the modulo algorithm, and floor represents the behavior of integer division.

[0037] Furthermore, in step (2),

[0038] Initialize the population using the A* algorithm: randomly select the starting coordinates and the ending coordinates, and add the starting coordinates and the ending coordinates to the chromosome encoding; add an obstacle-free random point as the target point in the underwater environment grid map model; use the A* algorithm to establish the paths from the starting coordinate point to the target point and from the target point to the ending coordinate point respectively, avoiding obstacle points and the peripheral area during the path establishment process; connect the starting coordinate point and the ending coordinate point, delete the path nodes that are exactly the same, and finally generate a path set;

[0039] Design the path fitness function: sample each path point in the path set, calculate the distance from each path point to the nearest obstacle point. When this distance is less than the value P, the fitness of the path increases; when this distance is greater than or equal to the value P, the fitness of the path decreases, where P is greater than or equal to half of the grid side length of the underwater environment grid map model; select the path with the minimum fitness as the good path:

[0040] Fitness(i) = ω1Length(p) + ω2Num

[0041] Among them, Fitness(i) is the fitness function, Length(p) is the path length, Num is the number of path points on the path whose distance to the obstacle point is less than P; ω1, ω2 are the proportionality coefficients of the path length and the number of path points close to the obstacle path. When ω1 > ω2, it means that ensuring a shorter path length is more prioritized. When ω1 < ω2, it means that staying away from obstacles to ensure operation safety is more prioritized;

[0042] Perform crossover and mutation operations: The individuals are paths, and one individual is a path. Perform crossover and mutation operations to generate more offspring individuals;

[0043] The crossover probability is P cIt is shown that the crossover operation in path planning refers to swapping two parent individuals at the intersection point after they have been searched, and after the swap, the shorter path is retained while the longer path is discarded; after the crossover operation, the fitness of the offspring individual is lower than that of the parent individual.

[0044] The mutation probability is denoted as P m It is shown that the mutation operation in path planning refers to flipping the searched parent individual with probability P m and then combining with the crossover operation to obtain an offspring individual with lower fitness.

[0045] The adaptive model of the crossover probability and the mutation probability is as follows:

[0046]

[0047]

[0048] where P c and P m are the crossover probability and the mutation probability respectively; are the maximum crossover probability and the minimum crossover probability of each generation, are the maximum mutation probability and the minimum mutation probability of each generation; f, f avg , f max and f min are the fitness function value, the average fitness function value, the maximum value of the average fitness function value, and the minimum value of the average fitness function value of each generation respectively;

[0049] When the fitness of an individual is lower than the average fitness, the crossover and mutation probabilities of the individual remain at the maximum value When the fitness of an individual is higher than the average fitness, the higher the fitness, the lower the crossover and mutation probabilities;

[0050] Perform a selection operation on the path: Select the most elite individual with the lowest fitness in the parent population as the good path, retain the good path for comparison with the most elite individual in the offspring population. If the fitness of the most elite individual in the offspring is less than the fitness of the good path, then update the most elite individual in the offspring to the good path; otherwise, still retain the most elite individual in the parent as the good path.

[0051] Furthermore, in step (3),

[0052] Assume that there is a Bezier curve containing m control points P0, P1, P2, …, P m The mathematical expression of this curve is as follows:

[0053]

[0054] where m ∈ N+, t is the normalized time variable, Pi =(x i , y i ) T is the coordinate vector of the i-th control point, where x i and y i are the components of the X and Y coordinates respectively, is the Bernstein polynomial, which is the basic function of the Bezier curve expression. The expansion is:

[0055] The above mathematical expression of the Bezier curve is used to optimize the good path to obtain the optimal path.

[0056] The present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned search and rescue ROV control method are implemented.

[0057] The present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned search and rescue ROV control method are implemented.

[0058] Beneficial effects: Compared with the prior art, the present invention has the following remarkable advantages: 1. It can stay away from obstacles: a fitness function is constructed using the path length and the distance from the path node to the nearest obstacle to achieve the purpose of staying away from obstacles; 2. It has a better initial path: in the improved genetic algorithm, a better initial path is obtained by introducing the A* algorithm; 3. It has a path with better quality and smoothness: by introducing the Bezier optimization operator and taking advantage of the continuity and smoothness of the Bezier curve, the quality and smoothness of the path are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is the flowchart of the present invention;

[0060] Figure 2 is the underwater environment grid map model diagram;

[0061] Figure 3 is the flowchart of path planning using the improved genetic algorithm and the Bezier optimization operator;

[0062] Figure 4 is the flowchart of the path fitness function;

[0063] Figure 5 is the block diagram of the underwater robot control system;

[0064] Figure 6 is the functional architecture diagram of path planning;

[0065] Figure 7It is a flowchart for automatically planning a path. Detailed implementation manner

[0066] The following will describe the detailed implementation manner of the present invention in conjunction with the accompanying drawings.

[0067] Embodiment 1

[0068] As Figure 1 shown, the search and rescue ROV control system includes:

[0069] An underwater environment grid map model establishment module for constructing an underwater environment grid map model through the working environment of the underwater robot.

[0070] As Figure 2 shown, in the underwater environment grid map model, the black grid represents an obstacle, and the white grid represents the free space of the underwater robot. The grid is marked using the serial number marking method, and the numbers are incremented by 1 in sequence according to the principle from left to right and from bottom to top; in the underwater environment grid map model, S, 2, 3,..., E are grid nodes, and each grid node is a path point; among them, S is the starting position of the underwater robot, and E is the ending position of the underwater robot; the relationship between the given grid serial number T(i, j) and the rectangular coordinate point P(i, j)[X(i, j), Y(i, j)] is as follows:

[0071]

[0072] Among them, mod represents the behavior of the modulo algorithm, and floor represents the behavior of integer division.

[0073] As Figure 3 shown, the path planning module is used to plan a good path by adopting an improved genetic algorithm. The process of planning the path by the improved genetic algorithm is as follows: initializing the population using the A* algorithm, designing a path fitness function, and performing crossover, mutation, and selection operations on the path.

[0074] Initializing the population using the A* algorithm: randomly selecting the starting coordinate and the ending coordinate, and adding the starting coordinate and the ending coordinate to the chromosome encoding; adding an obstacle-free random point as the target point in the underwater environment grid map model; using the A* algorithm to establish the paths from the starting coordinate point to the target point and from the target point to the ending coordinate point respectively, and avoiding obstacle points and the peripheral area during the path establishment process; connecting the starting coordinate point and the ending coordinate point, deleting the paths with exactly the same path nodes, and finally generating a path set.

[0075] As Figure 4As shown in the figure, design the path fitness function: sample each path point in the path set, calculate the distance from each path point to the nearest obstacle point. When the distance is less than the value P, the fitness of the path increases; when the distance is greater than or equal to the value P, the fitness of the path decreases, where P is greater than or equal to half of the grid side length of the underwater environment grid map model; select the path with the minimum fitness as the good path:

[0076] Fitness(i) = ω1Length(p) + ω2Num

[0077] where Fitness(i) is the fitness function, Length(p) is the path length, and Num is the number of path points on the path whose distance to the obstacle point is less than P; ω1 and ω2 are the proportionality coefficients of the path length and the number of path points close to the obstacle. When ω1 > ω2, it means that ensuring a shorter path length is more prioritized. When ω1 < ω2, it means that staying away from the obstacle to ensure operation safety is more prioritized.

[0078] Perform crossover and mutation operations on the path: The individual is the path, and one individual is a path. Perform crossover and mutation operations to generate more offspring individuals.

[0079] The crossover probability is represented by P c In path planning, the crossover operation refers to exchanging the two parent individuals that have been searched at the intersection point, and keeping the shorter path and discarding the longer path after the exchange; after the crossover operation, the fitness of the offspring individual is lower than that of the parent individual.

[0080] The mutation probability is represented by P m In path planning, the mutation operation refers to flipping the searched parent individual with probability P m and then combining with the crossover operation to obtain an offspring individual with a lower fitness.

[0081] The adaptive model of the crossover probability and the mutation probability is:

[0082]

[0083]

[0084] where P c and P m are the crossover probability and the mutation probability respectively; are the maximum crossover probability and the minimum crossover probability of each generation respectively, are the maximum mutation probability and the minimum mutation probability of each generation respectively; f is the fitness function value of each generation, f avg is the average fitness function value, f max is the maximum value of the average fitness function value, f minThe minimum value of the average fitness function.

[0085] When the fitness of an individual is lower than the average fitness, the crossover and mutation probability of the individual remains at the maximum value When the fitness of an individual is higher than the average fitness, the greater the fitness, the lower the crossover and mutation probability.

[0086] Perform a selection operation on the path: Select the most elite individual with the minimum fitness in the parent population as the good path, and retain the good path for comparison with the most elite individual in the offspring population. If the fitness of the offspring most elite individual is less than the fitness of the good path, then update the offspring most elite individual to the good path; otherwise, still retain the parent most elite individual as the good path.

[0087] Such as Figure 3 As shown, the path optimization module is used to optimize the good path using the Bezier optimization operator to obtain the optimal path.

[0088] Assume there is a Bezier curve containing m control points P0, P1, P2,..., P m The mathematical expression of this curve is as follows:

[0089]

[0090] where m ∈ N+, t is the normalized time variable, P i =(x i , y i ) T is the coordinate vector of the i-th control point, x i and y i are the components of the X and Y coordinates respectively, is the Bernstein polynomial, which is the basic function of the Bezier curve expression, and the expansion is:

[0091] Use the above mathematical expression of the Bezier curve to optimize the good path to obtain the optimal path.

[0092] The control module is used to control the underwater robot to conduct search and rescue through the optimal path.

[0093] Such as Figure 5 As shown, the underwater robot control system is divided into a surface control system and an underwater control system.

[0094] The water surface control system includes a power supply system and a monitoring system. One end of the water surface power supply box is connected to the umbilical cable, which is responsible for providing the power required for the movement of the underwater robot body. The other end is connected to the monitoring computer through a network cable. The built-in router is responsible for connecting the lower computer and the upper computer. The monitoring computer, as a terminal, displays the current attitude, depth, temperature and humidity, pressure, voltage and current and other information of the underwater robot body in real time. At the same time, the robot can choose two control methods: handle control and path planning (automatic control).

[0095] The underwater control system includes an energy control system, a motion control system, a video observation system, a safety monitoring system and an information acquisition system. Among them, the energy control system is responsible for power transmission, supplying power to different modules, and is also responsible for the voltage conversion circuit and the drive circuit. The motion control system drives the motor to rotate forward and backward to control the movement of the underwater robot. The video observation system includes a camera (pan-tilt camera) and an underwater light. The camera is responsible for observing the surrounding environment, and when the environment is dark, the underwater light is turned on to assist lighting. The safety monitoring system is responsible for real-time collection of alarm module information in the electronic pressure-resistant cabin, and issues an alarm when a failure occurs. The information acquisition system is responsible for real-time collection of various sensor data and GPS data in the pressure-resistant cabin of the underwater robot and uploading them to the upper computer monitoring system.

[0096] As Figure 6 shown, the path planning function needs to be jointly realized by the water surface monitoring system and the underwater control system.

[0097] As Figure 7 shown, the path planning is responsible for automatically planning the path of the underwater robot. The process is as follows:

[0098] 1. Open the path planning window and switch to the automatic planning mode.

[0099] 2. Environmental information processing. First, obtain the map model of the current sea area, set the no-go area on the map model, and perform image dilation and erosion processing to obtain the final grid map model.

[0100] 3. Select the target point of path planning, and use the current position of the underwater robot as the starting point of path planning and the target position as the end point.

[0101] 4. Judge whether the selected target point is reasonable. If it is not reasonable, reselect it.

[0102] 5. Call the path planning interface introducing the A* algorithm, input the map model, starting point, and target point, and perform path planning.

[0103] 6. Obtain the planned path nodes and perform task planning.

[0104] Embodiment 2

[0105] As Figure 1As shown in the figure, the search and rescue ROV control method includes:

[0106] (1) Construct an underwater environment grid map model through the working environment of the underwater robot.

[0107] As Figure 2 shown in the figure, in the underwater environment grid map model, the black grid represents an obstacle, and the white grid represents the free space of the underwater robot. The grid is marked using the serial number marking method, and the numbers are incremented by 1 in sequence according to the principle from left to right and from bottom to top; in the underwater environment grid map model, S, 2, 3,..., E are grid nodes, and each grid node is a path point; among them, S is the starting position of the underwater robot, and E is the ending position of the underwater robot; the relationship between the given grid serial number T(i, j) and the rectangular coordinate point P(i, j)[X(i, j), Y(i, j)] is as follows:

[0108]

[0109] Among them, mod represents the behavior of the modulo algorithm, and floor represents the behavior of integer division.

[0110] (2) As Figure 3 shown in the figure, a good path is obtained by planning the path using an improved genetic algorithm. The process of planning the path using the improved genetic algorithm is as follows: Initialize the population using the A* algorithm, design a path fitness function, and perform crossover, mutation, and selection operations on the path.

[0111] Initialize the population using the A* algorithm: Randomly select the starting coordinate and the ending coordinate, and add the starting coordinate and the ending coordinate to the chromosome encoding; Add an obstacle-free random point as the target point in the underwater environment grid map model; Use the A* algorithm to establish the paths from the starting coordinate point to the target point and from the target point to the ending coordinate point respectively, and avoid obstacle points and the peripheral area during the path establishment process; Connect the starting coordinate point and the ending coordinate point, and delete the paths with identical path nodes. Finally, generate a path set.

[0112] As Figure 4 shown in the figure, design a path fitness function: Sample each path point in the path set, calculate the distance from each path point to the nearest obstacle point. When the distance is less than the value P, the fitness of the path increases; when the distance is greater than or equal to the value P, the fitness of the path decreases, where P is greater than or equal to half of the grid side length of the underwater environment grid map model; Select the path with the minimum fitness as the good path:

[0113] Fitness(i) = ω1Length(p) + ω2Num

[0114] Among them, Fitness(i) is the fitness function, Length(p) is the path length, and Num is the number of path points on the path whose distance from the obstacle point is less than P; ω1 and ω2 are the proportionality coefficients of the path length and the number of path points close to the obstacle. When ω1 > ω2, it means that ensuring a shorter path length is more prioritized. When ω1 < ω2, it means that staying away from the obstacle to ensure safe operation is more prioritized.

[0115] Perform crossover and mutation operations on the paths: The individuals are paths, and one individual is a single path. Crossover and mutation operations are carried out to generate more offspring individuals.

[0116] The crossover probability is represented by P c In path planning, the crossover operation refers to swapping the two parent individuals found at the intersection point. After the swap, the shorter path is retained and the longer path is discarded; after the crossover operation, the fitness of the offspring individual is lower than that of the parent individual.

[0117] The mutation probability is represented by P m In path planning, the mutation operation refers to flipping the found parent individual with a probability of P m and then combining with the crossover operation to obtain an offspring individual with a lower fitness.

[0118] The adaptive model of the crossover probability and the mutation probability is as follows:

[0119]

[0120]

[0121] Among them, P c and P m are the crossover probability and the mutation probability respectively; are the maximum crossover probability and the minimum crossover probability of each generation respectively, are the maximum mutation probability and the minimum mutation probability of each generation respectively; f is the fitness function value of each generation, f avg is the average fitness function value, f max is the maximum value of the average fitness function, f min is the minimum value of the average fitness function.

[0122] When the fitness of an individual is lower than the average fitness, the crossover and mutation probabilities of the individual remain at the maximum value When the fitness of an individual is higher than the average fitness, the greater the fitness, the lower the crossover and mutation probabilities.

[0123] Perform a selection operation on the path: Select the most elite individual with the smallest fitness in the parent population as the good path, and retain the good path for comparison with the most elite individual in the offspring population. If the fitness of the offspring most elite individual is less than the fitness of the good path, update the offspring most elite individual to the good path; otherwise, still retain the parent most elite individual as the good path.

[0124] (3) As Figure 3 shown, use the Bezier optimization operator to optimize the good path to obtain the optimal path.

[0125] Assume there is a Bezier curve containing m control points P0, P1, P2, …, P m . The mathematical expression of this curve is as follows:

[0126]

[0127] where m ∈ N+, t is the normalized time variable, P i = (x i , y i ) T is the coordinate vector of the i-th control point, and x i and y i are the components of the X and Y coordinates respectively. is the Bernstein polynomial, which is the basic function of the Bezier curve expression, and its expansion is:

[0128] Use the above mathematical expression of the Bezier curve to optimize the good path to obtain the optimal path.

[0129] (4) Control the underwater robot to conduct search and rescue through the optimal path.

[0130] As Figure 5 shown, the underwater robot control system is divided into a surface control system and an underwater control system.

[0131] The surface control system includes a power supply system and a monitoring system. One end of the surface power supply box is connected to the umbilical cable, which is responsible for providing the power required for the movement of the underwater robot body. The other end is connected to the monitoring computer through a network cable, and the built-in router is responsible for connecting the lower computer and the upper computer; the monitoring computer, as a terminal, displays the current attitude, depth, temperature and humidity, pressure, voltage and current and other information of the underwater robot body in real time. At the same time, the robot can select two control methods: joystick control and path planning (automatic control).

[0132] The underwater control system includes an energy control system, a motion control system, a video observation system, a safety monitoring system, and an information acquisition system. Among them, the energy control system is responsible for power transmission to supply power to different modules, and is also responsible for the voltage conversion circuit and the drive circuit; the motion control system drives the motor to rotate forward and backward to control the movement of the underwater robot; the video observation system includes a camera (pan-tilt camera) and an underwater light. The camera is responsible for observing the surrounding environment, and when the environment is dark, the underwater light is turned on to assist lighting; the safety monitoring system is responsible for real-time collection of alarm module information in the electronic pressure-resistant cabin and sending an alarm when a fault occurs; the information acquisition system is responsible for real-time collection of various sensor data and GPS data in the pressure-resistant cabin of the underwater robot and uploading them to the upper computer monitoring system.

[0133] As Figure 6 shown, the path planning function needs to be jointly realized by the surface monitoring system and the underwater control system.

[0134] As Figure 7 shown, the path planning is responsible for automatically planning the path of the underwater robot, and the process is as follows:

[0135] 1. Open the path planning window and switch to the automatic planning mode.

[0136] 2. Environmental information processing. First, obtain the map model of the current sea area, set the no-go area on the map model, and perform image dilation and erosion processing to obtain the final grid map model.

[0137] 3. Select the target point of path planning, and use the current position of the underwater robot as the starting point of path planning and the target position as the end point.

[0138] 4. Judge whether the selected target point is reasonable. If it is not reasonable, reselect it.

[0139] 5. Call the path planning interface introducing the A* algorithm, input the map model, starting point, and target point, and perform path planning.

[0140] 6. Obtain the planned path nodes and perform task planning.

[0141] The present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned search and rescue ROV control method are implemented.

[0142] The present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned search and rescue ROV control method are implemented.

Claims

1. A search and rescue ROV control system with path planning function, characterized in that, Including: An underwater environment grid map model building module for building an underwater environment grid map model through the working environment of an underwater robot; A path planning module for planning a good path by using an improved genetic algorithm. The process of planning a path by the improved genetic algorithm is as follows: initializing the population by using the A* algorithm, designing a path fitness function, and performing crossover, mutation, and selection operations on the path; In the path planning module, Initializing the population by using the A* algorithm: randomly selecting a starting coordinate and an ending coordinate, and adding the starting coordinate and the ending coordinate to the chromosome encoding; adding an obstacle-free random point as a target point to the underwater environment grid map model; using the A* algorithm to establish paths from the starting coordinate point to the target point and from the target point to the ending coordinate point respectively, and avoiding obstacle points and peripheral areas during the path establishment process; Connecting the starting coordinate point and the ending coordinate point, deleting the path nodes that are exactly the same, and finally generating a path set; Designing a path fitness function: sampling each path point in the path set, calculating the distance from each path point to the nearest obstacle point, increasing the fitness of the path when the distance is less than the value P, and decreasing the fitness of the path when the distance is greater than or equal to the value P, where P is greater than or equal to half of the grid side length of the underwater environment grid map model; selecting the path with the minimum fitness as the good path: Fitness(i) = ω1Length(p) + ω2Num where Fitness(i) is the fitness function, Length(p) is the path length, Num is the number of path points on the path whose distance to the obstacle point is less than P; ω1 and ω2 are the proportionality coefficients of the path length and the number of path points close to the obstacle path. When ω1 > ω2, it means that ensuring a shorter path length is more prioritized. When ω1 < ω2, it means that staying away from obstacles to ensure operation safety is more prioritized; Performing crossover and mutation operations on the path: the individual is the path, and one individual is a path, performing crossover and mutation operations to generate more offspring individuals; The crossover probability is denoted by P c which means that in path planning, the crossover operation refers to swapping two searched parent individuals at the intersection point, retaining the shorter path and discarding the longer path after the swap; after the crossover operation, the fitness of the offspring individuals is lower than that of the parent individuals. The mutation probability is represented by P m In path planning, the mutation operation means that the parent individuals that have been searched are flipped with probability P m and then combined with the crossover operation to obtain offspring individuals with lower fitness; The adaptive model of the crossover probability and the mutation probability is: Among them, P c and P m are the crossover probability and the mutation probability respectively; are the maximum crossover probability and the minimum crossover probability of each generation respectively, are the maximum mutation probability and the minimum mutation probability of each generation respectively; f is the fitness function value of each generation, f avg is the average fitness function value, f max is the maximum value of the average fitness function value, f min is the minimum value of the average fitness function value; When the fitness of an individual is lower than the average fitness, the crossover and mutation probability of the individual remains at the maximum value When the fitness of an individual is higher than the average fitness, the greater the fitness, the lower the crossover and mutation probability; Performing selection operation on the path: selecting the most elite individual with the minimum fitness in the parent population as the good path, retaining the good path for comparison with the most elite individual in the offspring population. If the fitness of the offspring most elite individual is less than the fitness of the good path, then updating the offspring most elite individual as the good path; otherwise, still retaining the parent most elite individual as the good path; A path optimization module for optimizing the good path by using a Bezier optimization operator to obtain the optimal path; A control module for controlling the underwater robot to conduct search and rescue through the optimal path.

2. The search and rescue ROV control system with path planning function according to claim 1, characterized in that: In the underwater environment grid map model building module, In the underwater environment grid map model, black grids represent obstacles, and white grids represent the free space of the underwater robot. The grid is marked using the serial number marking method, and the numbers are incremented by 1 in sequence according to the principle from left to right and from bottom to top. In the underwater environment grid map model, S, 2, 3, …, E are grid nodes, and each grid node is a path point. Among them, S is the starting position of the underwater robot, and E is the ending position of the underwater robot. The relationship between the given grid serial number T(i, j) and the rectangular coordinate point P(i, j)[X(i, j), Y(i, j)] is as follows: Among them, mod represents the behavior of the modulo algorithm, and floor represents the behavior of integer division.

3. The search and rescue ROV control system with path planning function according to claim 1, characterized in that: In the path optimization module, Assume that there is a Bézier curve containing m control points P0, P1, P2, …, P m , and the mathematical expression of this curve is as follows: where \(m\in N^+\), \(t\) is the normalized time variable, \(P\) n =(x n ,y n ) T is the coordinate vector of the \(n\)th control point, \(x\) n and \(y\) n are the components of the \(X\) and \(Y\) coordinates respectively, is the Bernstein polynomial, which is the basic function of the Bezier curve expression, and its expansion is: Use the above mathematical expression of the Bezier curve to optimize the good path to obtain the optimal path.

4. A control method for a search and rescue ROV with path planning function, characterized in that, Including: (1) Construct an underwater environment grid map model through the working environment of the underwater robot; (2) Plan the path by using an improved genetic algorithm to obtain a good path. The process of the improved genetic algorithm for path planning is as follows: Initialize the population using the A* algorithm, design a path fitness function, and perform crossover, mutation, and selection operations on the path; Initialize the population using the A* algorithm: Randomly select the starting coordinate and the ending coordinate, and add the starting coordinate and the ending coordinate to the chromosome encoding. Add an obstacle-free random point as the target point in the underwater environment grid map model. Use the A* algorithm to establish the paths from the starting coordinate point to the target point and from the target point to the ending coordinate point respectively, and avoid obstacle points and the peripheral area during the path establishment process; Connect the starting coordinate point and the ending coordinate point, delete the path nodes that are exactly the same, and finally generate a path set; Design a path fitness function: Sample each path point in the path set, calculate the distance from each path point to the nearest obstacle point. When the distance is less than the value P, the fitness of the path increases; when the distance is greater than or equal to the value P, the fitness of the path decreases, where P is greater than or equal to half of the grid side length of the underwater environment grid map model. Select the path with the minimum fitness as the good path: Fitness(i) = ω1Length(p) + ω2Num Among them, Fitness(i) is the fitness function, Length(p) is the path length, Num is the number of path points on the path whose distance to the obstacle point is less than P; ω1, ω2 are the proportionality coefficients of the path length and the number of path points close to the obstacle. When ω1 > ω2, it means that ensuring a shorter path length is more prioritized. When ω1 < ω2, it means that staying away from obstacles to ensure operation safety is more prioritized; Perform crossover and mutation operations: The individuals are paths, and one individual is a path. Perform crossover and mutation operations to generate more offspring individuals; The crossover probability is denoted by P c In path planning, the crossover operation means that two searched parent individuals are exchanged at the intersection point. After the exchange, the shorter path is retained and the longer path is discarded. After the crossover operation, the fitness of the offspring individual is lower than that of the parent individual. The mutation probability is denoted as P m In path planning, the mutation operation means that the searched parent individual is flipped with probability P m and then combined with the crossover operation to obtain an offspring individual with lower fitness The adaptive model of the crossover probability and the mutation probability is: Among them, P c and P m are the crossover probability and the mutation probability respectively; is the maximum crossover probability and the minimum crossover probability of each generation, is the maximum mutation probability and the minimum mutation probability of each generation; f, f avg , f max and f min are the fitness function value, the average fitness function value, the maximum value of the average fitness function value, and the minimum value of the average fitness function value of each generation respectively; When the fitness of an individual is lower than the average fitness, the crossover and mutation probability of the individual remains at the maximum value When the fitness of an individual is higher than the average fitness, the greater the fitness, the lower the crossover and mutation probability; Perform a path selection operation: Select the most elite individual with the smallest fitness in the parent population as the good path, and retain the good path for comparison with the most elite individual in the offspring population. If the fitness of the most elite individual in the offspring is less than the fitness of the good path, update the most elite individual in the offspring to the good path; otherwise, still retain the most elite individual in the parent as the good path; (3) Optimize the good path using the Bezier optimization operator to obtain the optimal path; (4) Control the underwater robot to conduct search and rescue through the optimal path.

5. The search and rescue ROV control method with path planning function according to claim 4, characterized in that: In step (1), In the underwater environment grid map model, black grids represent obstacles, and white grids represent the free space of the underwater robot. The grid is marked using the serial number marking method, and the numbers are incremented by 1 in sequence according to the principle from left to right and from bottom to top. In the underwater environment grid map model, S, 2, 3,..., E are grid nodes, and each grid node is a path point; among them, S is the starting position of the underwater robot, and E is the ending position of the underwater robot; the relationship between the given grid serial number T(i, j) and the rectangular coordinate point P(i, j)[X(i, j), Y(i, j)] is as follows: Among them, mod represents the behavior of the modulo algorithm, and floor represents the behavior of integer division.

6. The search and rescue ROV control method with path planning function according to claim 4, characterized in that: In step (3), Assume that there is a Bézier curve containing m control points P0, P1, P2, …, P m , and the mathematical expression of this curve is as follows: where \(m\in N^+\), \(t\) is a normalized time variable, \(P\) n =(x n ,y n ) T is the coordinate vector of the \(n\)th control point, and \(x\) n and \(y\) n are the components of the \(X\) and \(Y\) coordinates respectively, is the Bernstein polynomial, which is the basic function of the Bezier curve expression, and its expansion is: Optimize the good path using the above mathematical expression of the Bezier curve to obtain the optimal path.

7. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 4 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 4 to 6.

Images