Path planning method for aquaculture inspection robot

Abstract

The present invention provides a path planning method for an aquaculture inspection robot. The path length cost is constructed by the inspection path length of the inspection robot. The center position of each aquaculture tank is used as the obstacle center, and the obstacle influence radius is set. The obstacle cost is constructed based on the distance between the inspection robot and the obstacle. The narrow channel cost is constructed by the spacing between the path nodes of the inspection robot. A slippery area is set around the aquaculture tank, and a slippery environment cost is constructed based on the distance between the inspection robot's path and the slippery area. The fitness function is constructed by integrating , , and through the weight coefficient. Based on the fitness function, the path planning is performed through the ant colony algorithm. Based on the comprehensive evaluation of the path length and the obstacle threat, the system can dynamically adjust the path points to ensure the navigation accuracy and stability of the robot in a complex aquaculture environment.

Description

Technical Field

[0001] The present invention relates to the technical field of robot path planning, and in particular to a path planning method for an aquaculture inspection robot. Background Art

[0002] Factory aquaculture, such as recirculating aquaculture systems (RAS), is a modern approach that requires efficient resource utilization. These systems require regular monitoring to ensure stable operation and maintain optimal environmental conditions to protect fish health. To reduce manual labor and improve efficiency, intelligent inspection robots have been developed. However, the navigation of robots in these environments poses challenges due to the presence of multiple obstacles, narrow passages, and irregular structural layouts.

[0003] Traditional path planning algorithms, such as A* and Dijkstra algorithms, are suitable for structured and known environments, but they perform poorly in large-scale and dynamic environments due to low computational efficiency and poor adaptability. Meta-heuristic algorithms such as particle swarm optimization (PSO) and genetic algorithms (GA) have certain flexibility, but usually have a long convergence time and are prone to local optimality. Summary of the invention

[0004] The present invention proposes a path planning method for an aquaculture inspection robot to solve the technical problem of path planning in an environment with multiple obstacles and irregular layout.

[0005] In order to solve the above technical problems, the present invention provides a path planning method for an aquaculture inspection robot, comprising the following steps:

[0006] Step S1: Construct the path length cost by the inspection path length of the inspection robot ;

[0007] Step S2: Take the center position of each aquaculture tank as the obstacle center, set the obstacle influence radius, and construct the obstacle cost based on the distance between the inspection robot and the obstacle. ;

[0008] Step S3: Construct the channel narrowness cost by checking the distance between the path nodes of the robot ;

[0009] Step S4: Set a slippery area around the aquaculture tank, and construct a slippery environment cost based on the path of the inspection robot and the distance between the slippery area and the slippery area. ;

[0010] Step S5: Use the weight coefficient to , , and Perform integration and construct fitness function;

[0011] Step S6: Based on the fitness function, path planning is performed using an ant colony algorithm.

[0012] Preferably, the path length cost The expression is:

[0013] ;

[0014] In the formula, n Indicates the number of nodes on the path; Indicates i The coordinates of the nodes.

[0015] Preferably, the obstacle cost The expression is:

[0016] ;

[0017] In the formula, It represents the threat cost amplification factor; n Indicates the number of nodes on the path; Indicates the number of aquaculture tanks; Indicates i The coordinates of the nodes; Indicates j The center of the aquaculture tank; Indicates the obstacle influence radius;

[0018] Preferably, the channel narrowing cost The expression is:

[0019] ;

[0020] In the formula, n Indicates the number of nodes on the path; Indicates i The coordinates of the nodes; Indicates setting a threshold.

[0021] Preferably, the slippery environment cost The expression is:

[0022] ;

[0023] In the formula, n Indicates the number of nodes on the path; Indicates i The coordinates of the nodes; represents the additional cost weight; m represents the number of slippery areas; The number of boundary points of the slippery area; Indicates r Slippery areas q The coordinates of the boundary points.

[0024] Preferably, the fitness function The expression is:

[0025] ;

[0026] In the formula, To control the weight of the path length cost; To control the weight of obstacle cost; is the weight of controlling the cost of channel narrowing; The weight of the cost of controlling slippery environment.

[0027] Preferably, in the path length cost A path smoothness penalty term is added to .

[0028] Preferably, the path smoothness penalty term The expression is:

[0029] ;

[0030] In the formula, represents the smoothness penalty coefficient.

[0031] Preferably, in step S6, the ant colony algorithm is optimized using an optimizer of the Newton-Raphson method.

[0032] Preferably, in the process of optimizing the ant colony algorithm using the optimizer of the Newton-Raphson method, the path position is updated by the following expression:

[0033] ;

[0034] ;

[0035] ;

[0036] In the formula, It is the temporary path position optimized by Newton-Raphson search rule and ant colony algorithm; Indicates the current location; represents normally distributed random numbers; Indicates the current worst position; Indicates the current best position; Indicates the current location; Indicates the random adjustment amplitude of the path position; and represents a random position; and Control the weights between different paths; Represents a random matrix of 1×dim, where dim is the search space dimension.

[0037] The beneficial effects of the present invention include at least: the present invention obtains the position and path information of the autonomous detection robot in the aquaculture environment in real time, and combines multi-sensor fusion technology to intelligently optimize the robot path planning. Based on the comprehensive evaluation of path length and obstacle threats, the system can dynamically adjust the path points to ensure the navigation accuracy and stability of the robot in complex aquaculture environments. In addition, as an additional technical feature, the path optimization strategy combining Newton-Raphson optimization and ant colony algorithm is adopted to improve the overall convergence speed and path quality, adapt to the autonomous detection needs in a variety of different aquaculture environments, and have efficient path optimization capabilities. It can effectively solve technical problems such as poor path smoothness and frequent local optimal problems in traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of a method flow chart of an embodiment of the present invention;

[0039] Figure 2 A schematic diagram of path planning based on obstacle threats according to an embodiment of the present invention;

[0040] Figure 3 A schematic diagram showing a comparison of robot inspection path planning based on narrow threat cost according to an embodiment of the present invention;

[0041] Figure 4 A schematic diagram showing a comparison of robot inspection path planning based on slippery environment cost according to an embodiment of the present invention;

[0042] Figure 5 A schematic diagram of a process of fusing an ant colony algorithm with a Newton-Raphson optimizer in an embodiment of the present invention;

[0043] Figure 6 It is a schematic diagram of the test comparison between the NRBO-ACO algorithm of the embodiment of the present invention and the existing algorithm under four benchmark functions;

[0044] Figure 7 A schematic diagram showing a performance comparison between the NRBO-ACO algorithm of an embodiment of the present invention and an existing algorithm;

[0045] Figure 8 It is a schematic diagram of convergence iteration analysis curves of the NRBO-ACO algorithm according to an embodiment of the present invention and the traditional ant colony algorithm. DETAILED DESCRIPTION

[0046] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings 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 belong to the protection scope of the present invention.

[0047] Example 1

[0048] like Figure 1 As shown, an embodiment of the present invention provides a path planning method for an aquaculture inspection robot, comprising the following steps:

[0049] Step S1: Construct the path length cost by the inspection path length of the inspection robot .

[0050] Specifically, the path length cost measures the total length of the path. Assume that the path consists of n nodes, and the coordinates of each node are ,in The range is 1 to n. The path length cost L is calculated by the following formula:

[0051] (1)

[0052] Step S2: Take the center position of each aquaculture tank as the obstacle center, set the obstacle influence radius, and construct the obstacle cost based on the distance between the inspection robot and the obstacle. .

[0053] Specifically, in the RAS environment, the fish tank is a large fixed obstacle that has a significant impact on path planning. In this embodiment, the center of each fish tank is used as the obstacle center, and the obstacle influence radius is set to According to the actual size of the fish tank, set Figure 2 As shown, the center point represents the center of the aquaculture tank, the orange area represents the radius of the aquaculture tank itself, and the light yellow area represents the radius of the obstacle influence set to , and then impose threat costs on the path nodes, expressed as:

[0054] (2)

[0055] in, is the threat cost amplification factor, The number of fish tanks is a node position on the path, the fish tank is the center of each fish tank. Orange represents the actual model. The path is invalid in the orange area. The impact radius of entering a fish tank When the fish is inside, its threat cost increases, driving the path away from the edge of the fish tank, thereby reducing the risk of collision.

[0056] Step S3: Construct the channel narrowness cost by checking the distance between the path nodes of the robot .

[0057] Specifically, in narrow channels, the distribution of path nodes is constrained by environmental restrictions and optimization objectives. Due to the limited channel width, path nodes can only be distributed along one-dimensional corridors in narrow spaces, and the spacing between adjacent nodes is small. The variation of this spacing is positively correlated with the channel width, so it is possible to determine whether it is in a narrow area by calculating the node spacing and setting a threshold. Although the initial path generation of the ant colony algorithm is random, providing the possibility of exploring a diverse solution space, as the iteration proceeds, pheromones gradually guide the path to converge in a more cost-effective direction. Especially in narrow channels, the algorithm tends to select path points restricted by boundaries, resulting in a denser distribution of nodes, thus reflecting the characteristics of narrow areas.

[0058] Narrow area determination: When generating the initial path, the channel width is detected by calculating the node spacing. Less than the set threshold , then mark the area as a "narrow passage".

[0059] Narrow cost increment: Increase the cost increment of the path in a narrow area , indicating that the driving difficulty increases, the expression is:

[0060] (3)

[0061] Schematic diagram of path planning optimized by narrow threat cost, such as Figure 3 As shown in the figure, the red color represents the traditional planned path, and the blue color represents the planned path of the present invention. It can be seen that the traditional path only focuses on the path length and does not consider the narrowness of the channel, which reduces the success rate of the inspection robot.

[0062] Step S4: Set a slippery area around the aquaculture tank and construct a slippery environment cost based on the inspection robot's path and the distance to the slippery area. .

[0063] Specifically, the friction of slippery roads is low, which can easily cause the robot to slip or deviate from the path. Therefore, the dynamic adjustment of the path cost due to slippery factors needs to be considered in path planning.

[0064] In this embodiment, the slippery area and the boundary points of the slippery area are set, and additional cost weights are applied. To increase the stability requirement of the path, the expression is:

[0065] (4)

[0066] In the formula, m represents the number of slippery areas; The number of boundary points of the slippery area; Indicates r Slippery areas q The coordinates of the boundary points.

[0067] The schematic diagram of the path planning optimized by the slippery environment cost is as follows Figure 4 As shown in the figure, red represents the traditional planned path, and blue represents the planned path of the present invention. It can be seen that the traditional path only focuses on the path length and does not consider the impact of the slippery environment, so that the path of the inspection robot passes through the boundary points of the slippery environment area, increasing the risk of slipping.

[0068] Step S5: Use the weight coefficient to , , and Perform integration and construct a fitness function.

[0069] Specifically, considering the factors of narrowness, slippery and obstacles, the comprehensive fitness function can control the influence of path length, obstacle threat, narrow passage and slippery area through weight coefficients. The fitness function formula is:

[0070] (5)

[0071] in, To control the importance of path length and smoothness, To control the weights of the threat costs of the fish tank and obstacles, the path is more cautious when approaching the fish tank area. To control the cost of narrow channels and reduce the complexity of paths in narrow areas, In order to control the cost weight of the slippery environment, ensure that the path avoids the slippery area as much as possible and increase the stability of the inspection robot.

[0072] Step S6: Based on the fitness function, path planning is performed using an ant colony algorithm.

[0073] Specifically, the ant colony algorithm is a heuristic search strategy based on a positive feedback mechanism. Its operating principle is as follows: In the initial stage of the algorithm, ants start from the starting point and randomly select the next target node while avoiding obstacles until they reach the target node. During this process, ants secrete pheromones along the way to enhance the path recognition ability of other ants. By continuously iterating and updating the path selection and pheromone concentration, the algorithm can eventually effectively find the optimal path.

[0074] At the moment In the traditional ant colony algorithm, ants are based on the current node The ant colony algorithm calculates and compares the transfer probability of each node based on the pheromone concentration and heuristic function of the surrounding feasible nodes. Then, the ant selects the node with the highest transfer probability as the next moving target. This process is iterated continuously until the ant reaches the destination. The transition probability of selecting the next node can be calculated by formula (6).

[0075] (6)

[0076] In the formula, Indicates time t path arrive pheromone concentration; Indicates that the next node belongs to the set of selectable nodes; Is the current path node The set of surrounding feasible nodes; and Represent the weights of controlling pheromone and heuristic value respectively; Indicates time t path arrive The inspiration value of .

[0077] Pheromone update: When an ant completes an iteration, the pheromone concentration on its path needs to be updated. The method for updating ant pheromones is as follows:

[0078] (7)

[0079] (8)

[0080] In the formula, is the pheromone volatility coefficient; is the total number of ants; is the pheromone intensity; is the length of the path that the ant has traveled, Represents ants Already passed the path ).

[0081] In the ant colony algorithm ACO, population initialization is the key first step. By generating an initial population matrix, this embodiment provides a diversified search basis for the algorithm, thereby increasing the possibility of finding the optimal path. Each individual represents a path composed of multiple nodes. The specific steps are as follows:

[0082] 1. Based on the given environment and obstacle information, randomly generate an initial population, where each individual represents a path.

[0083] 2. Each path consists of multiple nodes, which are randomly distributed on the map and meet the requirements of the starting point and the end point. The specific formula for generating the initial population is shown in formula (9):

[0084] (9)

[0085] in, Indicates that the nth individual is in dimension; lb and ub represent the lower and upper limits of the search space respectively; rand represents a random number in the range [0,1], and dim represents the dimension of the search space, that is, the number of variables in the problem. In this way, the value of each dimension of each individual is randomly generated within the upper and lower limits.

[0086] Example 2

[0087] For inspection robots in RAS environments, path smoothness is particularly important. This embodiment adds a smoothness penalty term to the path length cost. The path length cost in Example 1 is replaced so that the path avoids sharp turns in narrow areas. The expression is:

[0088] (10)

[0089] in is the smoothness penalty coefficient, is the smoothness penalty term, combined with the path length cost :

[0090] (11)

[0091] Example 3

[0092] Based on Example 1, this embodiment optimizes the ant colony algorithm in step S6 based on the Newton-Raphson optimizer NRBO fusion ant colony algorithm, such as Figure 5 shown.

[0093] Specifically, the optimizer NRBO based on the Newton-Raphson method comprehensively explores the search process through two rules: the Newton-Raphson search rule NRSR and the trap avoidance operator TAO. Through these improvements, the exploration ability of the improved ant colony algorithm is enhanced, and the convergence speed of finding a better position in the search space in the factory aquaculture environment is increased. In addition, the TAO mechanism helps the improved ant colony algorithm to avoid local optimal traps, ensuring the navigation robustness and efficiency of the detection robot in complex aquaculture environments. It mainly includes the following parts.

[0094] 1. Calculate the derivative

[0095] Calculate the derivatives of the path length cost and obstacle threat cost:

[0096] (12)

[0097] in represents the fitness function, is the path length cost, , , They represent the threat costs of different types of obstacles. , , , is the corresponding weight coefficient.

[0098] 2. Path location update

[0099] First, the Newton-Raphson search rule NRSR is used to update the path position. The specific formulas are shown in (13), (14) and (15):

[0100] (13)

[0101] (14)

[0102] (15)

[0103] in is the adjustment value for position update, indicating the update step size; Indicates the current worst position, Indicates the current best position is the current location, Indicates the random adjustment amplitude of the path position, It is a normally distributed random number to ensure randomness in the update process.

[0104] 3. Trap Avoidance Operator TAO

[0105] Whether to execute the trap avoidance operator TAO depends on the difference between the fitness function value of the current path position and the fitness function value of the current best path position. A threshold ζ is defined. In this embodiment, ζ=0.001 is set.

[0106] If the difference between the two is less than this threshold, the TAO operation is performed. Specifically, if , then perform TAO operation.

[0107] In order to improve the NRBO framework, this embodiment introduces the parameter , which is used to adjust the impact of different path positions on the update process. By combining the current best position and random positions The fitness value of can enhance the algorithm's exploration ability in the solution space. The expression is:

[0108] (16)

[0109] in Indicates the best position, Indicates the current location. and is a random position. and parameters To control the weights between different paths, the path position update is replaced by:

[0110] (17)

[0111] in It is the temporary path position optimized by Newton-Raphson search rule NRSR and ant colony algorithm.

[0112] Temporary location The effects of parameter optimization in the Newton-Raphson search rule (NRSR) and the ant colony algorithm are combined. This temporary position prepares for the subsequent trap avoidance operator (TAO) step, ensuring that the algorithm can effectively navigate the complex search space and avoid local optimality.

[0113] Use the trap avoidance operation TAO to avoid local optima:

[0114] (18)

[0115] (19)

[0116] (20)

[0117] in is the path location after the trap avoidance operation, and is the control parameter, and is based on The random factor calculated, The fitness impact of control parameters, is a random number between (0,1), Used to adjust the impact of the number of iterations, where is the current iteration number, is the maximum number of iterations, and the final path position is updated as:

[0118] (twenty one)

[0119] also, According to the fitness value of the new path position, the current optimal path position is updated through the following formula and the worst path location .

[0120] (twenty two)

[0121] (twenty three)

[0122] in is the fitness value of the new path position.

[0123] Figure 6 The figure shows the comparison between the algorithm optimized in this embodiment and the existing algorithm under four benchmark functions. , , and Four different benchmark functions are represented. It can be seen that the NRBO-ACO algorithm of the embodiment of the present invention is superior to other methods in terms of cost optimization and convergence speed, which shows its versatility and adaptability.

[0124] Figure 7 Table 1 shows the comparison between the NRBO-ACO path planning algorithm, the traditional ant colony algorithm and two other common optimization ant colony algorithms. Figure 7 (a) is the adaptive particle swarm-ant colony optimization algorithm. Figure 7 (b) is the traditional ant colony algorithm. Figure 7 (c) in the figure is the navigator-ant colony optimization algorithm. Figure 7 (d) in FIG. 1 is the NRBO-ACO algorithm of this embodiment. It can be seen that the NRBO-ACO algorithm of this embodiment has significant improvements in path smoothness, number of inflection points, and path length, especially showing higher robustness in complex environments.

[0125] Table 1

[0126]

[0127] Figure 8 The convergence iteration analysis curves of the NRBO-ACO algorithm and the traditional ant colony algorithm are shown. As can be seen from the figure, the NRBO-ACO algorithm of the red curve has obvious improvements in convergence speed and global optimization ability compared to the traditional ACO algorithm of the blue curve. The NRBO-ACO algorithm can achieve better path planning results with fewer iterations.

[0128] Example 4

[0129] In order to improve the adaptability and efficiency of the ant colony algorithm in solving optimization problems, this embodiment designs a dynamic parameter adjustment strategy. This strategy uses the iteration information during the algorithm execution process, including the current iteration number. and the maximum number of iterations , to dynamically adjust key parameter values , as shown in formula (24).

[0130] (twenty four)

[0131] In the ants' path selection process, the intensity of pheromones is regulated by specific parameters, which also determine the relative importance of heuristic information (such as path length). In addition, another parameter controls the volatility of pheromones. The core purpose of dynamically adjusting these parameters is to balance exploration and exploitation at different stages of algorithm execution, thereby accelerating the convergence process. The specific dynamic adjustment strategy is as follows:

[0132] 1. Adaptive adjustment of parameters: As the algorithm progresses, the preference for pheromones gradually increases, thereby making more effective use of the currently discovered optimal solution. The calculation formula is shown in formula (25):

[0133] (25)

[0134] 2. Adaptive adjustment of parameters: In the early stages of the algorithm, the influence of heuristic information is enhanced. As the algorithm progresses, its effect gradually weakens, while the weight of pheromone gradually increases. The calculation formula is shown in formula (26):

[0135] (26)

[0136] 3. Adaptive adjustment of parameters: In the later stage of the algorithm, it is particularly important to reduce the pheromone volatilization rate, which helps to maintain the concentration of pheromones and effectively promotes the convergence of the algorithm. The calculation formula is shown in formula (27):

[0137] (27)

[0138] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. Only the preferred embodiments of the present invention are expressed. The description is more specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. As long as there is no contradiction in the combination of these technical features, they should be considered as the scope recorded in this specification.

[0139] It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A path planning method for an aquaculture inspection robot, characterized in that: The following steps are involved: Step S1: Construct the path length cost by the inspection path length of the inspection robot ; Step S2: Take the center position of each aquaculture tank as the obstacle center, set the obstacle influence radius, and construct the obstacle cost based on the distance between the inspection robot and the obstacle. ; Step S3: Construct the channel narrowness cost by checking the distance between the path nodes of the robot ; Step S4: Set a slippery area around the aquaculture tank, and construct a slippery environment cost based on the path of the inspection robot and the distance between the slippery area and the slippery area. ; Step S5: Use the weight coefficient to , , and Perform integration and construct fitness function; Step S6: Based on the fitness function, path planning is performed using an ant colony algorithm.

2. The path planning method of an aquaculture inspection robot according to claim 1, characterized in that: The path length cost The expression is: ； In the formula, n Indicates the number of nodes on the path; Indicates i The coordinates of the nodes.

3. The path planning method of an aquaculture inspection robot according to claim 1, characterized in that: The obstacle cost The expression is: ； In the formula, It represents the threat cost amplification factor; n Indicates the number of nodes on the path; Indicates the number of aquaculture tanks; Indicates i The coordinates of the nodes; Indicates j The center of the aquaculture tank; Indicates the obstacle influence radius.

4. The path planning method for an aquaculture inspection robot according to claim 1, characterized in that: The cost of the narrow channel The expression is: ； In the formula, n Indicates the number of nodes on the path; Indicates i The coordinates of the nodes; Indicates setting a threshold.

5. The path planning method for an aquaculture inspection robot according to claim 1, characterized in that: The cost of slippery environment The expression is: ； In the formula, n Indicates the number of nodes on the path; Indicates i The coordinates of the nodes; represents the additional cost weight; m represents the number of slippery areas; The number of boundary points of the slippery area; Indicates r Slippery areas q The coordinates of the boundary points.

6. The path planning method for an aquaculture inspection robot according to claim 1, characterized in that: The fitness function The expression is: ； In the formula, To control the weight of the path length cost; To control the weight of obstacle cost; is the weight of controlling the cost of channel narrowing; The weight of the cost of controlling slippery environment.

7. The path planning method for an aquaculture inspection robot according to claim 2, characterized in that: The path length cost A path smoothness penalty term is added to .

8. The path planning method for an aquaculture inspection robot according to claim 7, characterized in that: The path smoothness penalty The expression is: ； In the formula, represents the smoothness penalty coefficient.

9. The path planning method for an aquaculture inspection robot according to claim 1, characterized in that: In step S6, the ant colony algorithm is optimized using a Newton-Raphson method optimizer.

10. The path planning method for an aquaculture inspection robot according to claim 9, characterized in that: In the process of optimizing the ant colony algorithm using the Newton-Raphson method optimizer, the path position is updated by the following expression: ； ； ； In the formula, It is the temporary path position optimized by Newton-Raphson search rule and ant colony algorithm; Indicates the current location; represents normally distributed random numbers; Indicates the current worst position; Indicates the current best position; Indicates the current location; Indicates the random adjustment amplitude of the path position; and represents a random position; and Control the weights between different paths; Represents a random matrix of 1×dim, where dim is the dimension of the search space.

Images