An underwater target searching path planning method and system based on an ant colony algorithm

Abstract

The application discloses an underwater target searching path planning method and system based on an ant colony algorithm, belongs to the technical field of ship and ocean engineering, and is used for planning an underwater target searching path and comprises the following steps: a heterogeneous grid graph is constructed by fusing target probability, water depth, obstacles and ocean currents; an independent ant colony is allocated to each underwater equipment, path pheromone and allocation pheromone are maintained, and the joint solution of regional segmentation, equipment allocation and path planning is realized through coupled state transition probability; an equipment task depth adaptation constraint is introduced to ensure that professional equipment is used for professional water depth; a meeting type bidirectional path construction strategy is adopted to accelerate convergence; a probability graph is updated through Bayes posterior according to real-time search feedback, and local re-planning is triggered. The application greatly improves cumulative discovery probability, convergence speed and resource adaptability, has the beneficial effects of strong dynamic response capability, zero deep violation of equipment, high coverage rate of high-probability areas and the like, and can be widely applied to underwater searching emergency tasks.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding and marine engineering technology, and more particularly to the field of path planning technology for underwater target search equipment, specifically to an underwater target search path planning method and system based on ant colony algorithm. Background Technology

[0002] After a target falls underwater, its actual location is highly uncertain due to ocean currents, topography, and positioning errors. Existing technologies typically generate a "probability map of the underwater landing location of the target" through drift trajectory inversion or acoustic beacon positioning to guide the search and rescue of underwater targets.

[0003] Currently, existing underwater search path planning technologies have the following drawbacks:

[0004] First, existing underwater search methods typically involve manually dividing the area into blocks and then selecting underwater search equipment based on past search experience within these blocks. This approach fails to deeply couple the "target probability distribution" with the "equipment performance parameters." This segmented optimization model leads to a loss of global optimality, resulting in a disconnect between area allocation and path planning. High-probability areas for underwater landing points may be occupied by inefficient search equipment, severely impacting search efficiency.

[0005] Secondly, HOVs (manned underwater vehicles) are suitable for detailed reconnaissance of key areas, AUVs (autonomous underwater robots) are suitable for large-scale general surveys, and ROVs (remotely operated vehicles) are suitable for precise exploration of fixed points. Existing technologies lack a unified mathematical (parametric) model for scheduling these heterogeneous underwater search equipment, and usually adopt a "first-come, first-served" or fixed partition mode, which leads to difficulties in coordinating heterogeneous equipment and a serious mismatch between equipment performance and mission requirements.

[0006] Third, existing technologies typically use the probability map of the underwater landing location of the crashed target as a one-time input. The information in the probability map is only used statically and the probability distribution is not dynamically updated based on real-time search feedback (no target detected). This causes underwater search equipment to repeatedly search low-value areas, resulting in a waste of resources.

[0007] Fourth, traditional ocean sweeping paths (such as zigzag and parallel line scanning) assume a uniform probability distribution in the search area, without considering the spatial heterogeneity of the probability map and the non-uniformity of underwater landing probability. This results in a large amount of search energy and time being wasted in low-probability areas, leading to insufficient search coverage density in high-probability areas.

[0008] Therefore, there is an urgent need to develop an underwater target search path planning method and system that simultaneously solves the problems of underwater heterogeneous equipment area division, task allocation, and three-dimensional path planning, so as to improve the precision measurement capability of near-infrared moisture detection devices in dynamic and changing scenarios. Summary of the Invention

[0009] The purpose of this invention is to provide an underwater target search path planning method and system based on ant colony algorithm, so as to solve the problems of low search efficiency for underwater crashed targets and difficulty in the coordinated operation of heterogeneous underwater search equipment in the prior art.

[0010] To achieve the above objectives, this invention provides an underwater target search path planning method based on ant colony algorithm, comprising:

[0011] S1. Construct a heterogeneous grid map, including obtaining a probability map of the underwater impact point of the crashed target, and integrating water depth topography data, seabed obstacle distribution data, and ocean current field data;

[0012] S2. Parametrically model each underwater equipment involved in the search and construct an underwater equipment cluster;

[0013] S3. Construct an ant colony equipment mapping mechanism, assign an independent ant colony to each underwater equipment, and each independent ant colony maintains the path pheromone and allocation pheromone of the corresponding equipment.

[0014] S4. Construct state transition rules and calculate the state transition probabilities of coupled path pheromones, allocated pheromones, and heterogeneous grid diagrams;

[0015] S5. Introduce equipment grid depth adaptation constraints to eliminate grids that underwater equipment cannot reach in heterogeneous grid maps;

[0016] S6. Adopting a bidirectional path construction strategy based on encounters, a reference search path is constructed based on constraint state transition rules and equipment grid depth adaptation constraints.

[0017] S7. Based on the reference search path provided by S6, search for the crashed target. During the search process, based on the search feedback from each underwater equipment, update the heterogeneous grid map of S1 through Bayesian posterior and output the search area boundary and optimal search path of each underwater equipment.

[0018] In S1, heterogeneous raster diagram for:

[0019] ;

[0020] In the formula, This represents the probability matrix indicating the underwater impact point of the crashed target. This represents the water depth matrix of the target sea area. This represents the obstacle occupancy matrix in the target area. This represents the ocean current intensity matrix for the target sea area, which is the sea area where the lost target is to be salvaged.

[0021] In S2, underwater equipment cluster for:

[0022] ;

[0023] In the formula, Indicates the first Parameter vectors of underwater equipment in Taiwan This indicates the total number of underwater equipment involved in the search.

[0024] In S3, for each Assign an independent ant colony each include Each ant colony independently maintains two pheromone matrices, including a path pheromone matrix and an allocation pheromone matrix.

[0025] In S4, the state transition rules, coupling path pheromone matrix, allocation pheromone matrix, and heterogeneous grid map are used to calculate the state transition probability using the following formulas:

[0026] ;

[0027] ;

[0028] ;

[0029] ;

[0030] In the formula, express time Ants in the grid cell The probability, express time Ants in the grid cell The path pheromone matrix, Represents grid cells heuristic function, Represents grid With grid The distance between them This indicates that the crashed target appeared in the grid. The probability in express time In the grid The allocation of pheromone matrix, This indicates that the ants are in the grid cells. The turning cost factor For underwater equipment heading and grid Pointing grid The angle between directions, This indicates that the ants are in the grid cells. Ocean current contributing factors, This indicates that the ants are in the grid cells. The angle between the course of the underwater equipment and the direction of the ocean current. , , , , These represent the weight parameters corresponding to path pheromones, heuristic functions, allocation pheromones, turning cost factors, and ocean current assist factors, respectively. Indicates the index of the candidate node. express The candidate node set, express time In grid cells Path pheromones, Represents grid cells heuristic function, express time In candidate nodes The distribution of pheromones, This indicates that the ants are in the grid cells. The turning cost factor This indicates that the ants are in the grid cells. Ocean currents are a contributing factor.

[0031] In S5, an equipment grid depth adaptation constraint is introduced to construct the equipment grid depth adaptation matrix:

[0032] ;

[0033] In the formula, Indicates the first Taiwan Underwater Equipment and Grid Unit Adaptability Represents grid The water depth, For the first The optimal diving depth for Taiwan's underwater equipment. For the first The maximum diving depth of the underwater equipment is determined by embedding the constructed equipment grid depth adaptation matrix as a multiplicative factor into the state transition probability calculation formula as a heuristic function.

[0034] In S6, based on the state transition rules after embedding equipment grid depth adaptation constraints, the ant colony algorithm is used to find the target search endpoint from the heterogeneous grid map. A meeting-based bidirectional path construction strategy is adopted, constructing paths from both the search start point and the target search endpoint simultaneously, and completing path splicing at the meeting point. The reference search path is output, and the fitness of ants in each independent ant colony is calculated. :

[0035] ;

[0036] In the formula, This represents the sum of the cumulative discovery probabilities of the path-covered raster. Represents the total probability over the entire domain. The latest time that underwater equipment can complete its search mission. Indicates the maximum allowed operation time for the search task. This indicates the total energy consumption of underwater equipment. This represents the expected energy consumption to complete the search task. This indicates the total depth of the operating grid beyond the maximum diving depth of the underwater equipment. This indicates the maximum allowable ultra-deep accumulation. , , , , These represent the weight coefficients of the corresponding items.

[0037] During the construction of reference search paths, each independent ant colony updates its pheromones based on its fitness value, including local updates of path pheromones and global updates of pheromone allocation.

[0038] The local update of path pheromones is based on the local update rules of the ant colony algorithm, whereby the path pheromones are updated after each ant completes its path construction:

[0039] ;

[0040] In the formula, express time Ants in the grid cell The path pheromone matrix represents the path pheromone matrix in the path pheromone matrix. Time grid unit pheromone concentration, Indicates the local pheromone increment. Represents the local update coefficients. This indicates the pheromone volatile term. This indicates the pheromone deposition term.

[0041] The global update of pheromone allocation follows a global update rule based on the MMAX (Maximum-Minimum Ant System) ant colony algorithm. After each independent ant colony constructs a complete reference search path, the fitness calculation formula is used to calculate... The fitness values ​​of each ant in the data are ranked first. Only a few ants receive an enhancement update, while the remaining ants receive a penalty decay update.

[0042] ;

[0043] ;

[0044] In the formula, express time Ants in the grid cell The allocation of pheromone matrix, Indicates the first All the grid cells visited by the ant Indicates the global update coefficient. Indicates the first The fitness of ants express The sum of the fitness of all ants is used to normalize the fitness of each ant.

[0045] The ant colony algorithm is used to search for the crashed target on the reference search path provided by S6. During the search process, based on the search feedback that no crashed target was detected, the probability of the underwater landing point of the crashed target in the heterogeneous grid image is updated using Bayesian posterior.

[0046] ;

[0047] In the formula, Indicates in During the ongoing search, the crashed target was located within a grid cell. The probability of; Indicates in During the ongoing search, the crashed target was located within a grid cell. The probability of; Indicates the probability of equipment detection;

[0048] Set iteration termination conditions, including the maximum number of iterations. Iterative improvement of limit When the algorithm iterations exceed the maximum number of iterations, or during the iteration process, continuous iterations... If the optimal solution does not improve, the iteration terminates and outputs the search area boundary of each underwater equipment, the optimal search path of each underwater equipment, and the expected cumulative discovery probability curve of the crashed target.

[0049] To achieve the above objectives, the present invention also provides an underwater target search path planning system based on ant colony algorithm, including a heterogeneous grid map construction module, a multi-ant colony cooperative control module, and a multi-equipment execution control module;

[0050] The heterogeneous grid map construction module includes multibeam echo sounder, side-scan sonar or synthetic aperture sonar, Doppler current profiler, satellite positioning system or beacon positioning system, used to obtain probability maps of the underwater landing points of crashed targets. In the obtained probability maps of the underwater landing points of crashed targets, water depth and topography data, seabed obstacle distribution data and ocean current field data of the target sea area are integrated to construct a heterogeneous grid map to guide the search for crashed targets in the target sea area.

[0051] The multi-ant colony collaborative control module includes an edge computing unit and an airborne computing unit, which are used to execute the ant colony algorithm to generate search path schemes. The edge computing unit is equipped with a three-layer heterogeneous pheromone matrix, including a path pheromone matrix, an allocation pheromone matrix, and an equipment grid depth adaptation matrix. The path pheromone matrix, the allocation pheromone matrix, and the heuristic function calculated based on the heterogeneous grid graph are coupled through a state transition probability calculation formula. The equipment grid depth adaptation matrix is ​​embedded in the state transition probability calculation formula as a multiplicative factor of the heuristic function. The airborne computing unit is deployed on the underwater equipment and is used to perform local path replanning and obstacle avoidance tasks.

[0052] The multi-equipment execution control module includes an equipment parameterization module, a path execution module, and an information feedback module. It is used to guide the actual search tasks of each underwater equipment according to the search path plan and to provide real-time feedback on the search results.

[0053] Multibeam echo sounders are used to acquire real-time water depth and topographic data of the target sea area; side-scan sonar or synthetic aperture sonar is used to acquire the distribution of obstacles in the target sea area; Doppler current profilers are used to acquire real-time three-dimensional ocean current field data of the target sea area; satellite positioning systems are used to obtain the probability map of the underwater landing point of the crashed target through drift trajectory inversion; and beacon positioning systems are used to obtain the probability map of the underwater landing point of the crashed target through acoustic beacon positioning.

[0054] The equipment parameterization module is used to perform parameterized modeling of the underwater equipment participating in the search, obtain the parameter vectors corresponding to each underwater equipment, and build an underwater equipment cluster based on each parameter vector. The path execution module is used to generate search path schemes according to the multi-ant colony collaborative control module, and schedule each underwater equipment to participate in the actual search task of underwater crashed targets. The information feedback module is used to transmit the search results of crashed targets in real time. When the transmitted search result is that no target was detected, Bayesian posterior probability decay is performed. When the transmitted search result is that suspected target features or equipment failure is found, local ant colony replanning is triggered.

[0055] Compared with the prior art, the present invention has the following advantages:

[0056] This invention employs a multi-ant colony collaborative architecture to jointly solve region segmentation, equipment allocation, and path planning, avoiding information loss during phased optimization. Simulation experiments show that within a 6-hour search window, the cumulative discovery probability of this invention reaches 79.8%, a 94% improvement over the traditional zigzag search (41.2%) and an 18.6% improvement over the standard ant colony algorithm (67.3%).

[0057] The proposed meeting-based bidirectional path construction strategy effectively solves the problem of slow convergence of ant colony algorithms in large-scale sea areas. In a typical scenario of 20 nautical miles × 20 nautical miles, the number of iterations required for convergence is reduced from 380 generations in the standard ant colony algorithm to 220 generations, improving the convergence speed by 42%.

[0058] This invention, by constructing an equipment-grid fit matrix and embedding it into state transition probabilities, fundamentally avoids resource waste and safety risks such as deep-water equipment idling and shallow-water equipment operating at excessive depths. In all simulation experiments, the number of instances of equipment violating regulations at excessive depths remained zero.

[0059] This invention uses a probability-weighted heuristic function and a pheromone-guided mechanism to automatically tilt search resources towards high-probability regions. The coverage of high-probability regions (the top 10% of grid cells by probability value) is increased from 53% in traditional methods to 94%, significantly reducing the risk of missed detections.

[0060] This invention introduces Bayesian posterior probability updates and association rule-driven updates, enabling the probability graph to dynamically evolve with real-time search feedback. When a suspected target feature is detected, the system can complete local replanning within 0.8 seconds and quickly dispatch the ROV to verify it, filling the gap in existing technologies that rely on "one-time probability graphs".

[0061] The method of this invention is not dependent on specific sea area characteristics or equipment models, and is applicable to underwater search scenarios of different scales, water depths, and formations. The system architecture adopts a modular design, allowing for flexible addition or reduction of equipment quantity according to actual mission requirements. Attached Figure Description

[0062] Figure 1 This is a probability distribution map of underwater accident targets provided by the present invention;

[0063] Figure 2 This is a diagram of the equipment allocation area provided by the present invention;

[0064] Figure 3 This is a search path diagram for various underwater devices provided by the present invention. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention are described clearly and completely below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0066] An underwater target search path planning method based on ant colony algorithm includes:

[0067] S1. Construct a heterogeneous grid map, including obtaining a probability map of the underwater impact point of the crashed target, and integrating water depth topography data, seabed obstacle distribution data, and ocean current field data;

[0068] S2. Parametrically model each underwater equipment involved in the search and construct an underwater equipment cluster;

[0069] S3. Construct an ant colony equipment mapping mechanism, assign an independent ant colony to each underwater equipment, and each independent ant colony maintains the path pheromone and allocation pheromone of the corresponding equipment.

[0070] S4. Construct state transition rules and calculate the state transition probabilities of coupled path pheromones, allocated pheromones, and heterogeneous grid diagrams;

[0071] S5. Introduce equipment grid depth adaptation constraints to eliminate grids that underwater equipment cannot reach in heterogeneous grid maps;

[0072] S6. Adopting a bidirectional path construction strategy based on encounters, a reference search path is constructed based on constraint state transition rules and equipment grid depth adaptation constraints.

[0073] S7. Based on the reference search path provided by S6, search for the crashed target. During the search process, based on the search feedback from each underwater equipment, update the heterogeneous grid map of S1 through Bayesian posterior and output the search area boundary and optimal search path of each underwater equipment.

[0074] In S1, a probability map of the underwater landing point location of the crashed target is generated by inverting existing drift trajectories or using acoustic beacon positioning. This map is then integrated with water depth and topographic data obtained from multibeam echo sounding, seabed obstacle distribution data obtained from side-scan sonar or synthetic aperture sonar, and three-dimensional ocean current field data obtained from an acoustic Doppler current profiler to construct a multidimensional heterogeneous grid map for guiding underwater searches. :

[0075] ;

[0076] In the formula, This represents the probability matrix indicating the underwater impact point of the crashed target. This represents the water depth matrix of the target sea area. This represents the obstacle occupancy matrix in the target area. This represents the ocean current intensity matrix for the target sea area, which is the sea area where the lost target is to be salvaged.

[0077] In S2, at least two types of underwater equipment are selected from intelligent underwater robots (AUVs), remotely operated underwater vehicles (ROVs), and manned underwater vehicles (HOVs). Performance parameters of each participating underwater vehicle are read, including maximum diving depth, endurance, cruising speed, swath width, turning radius, operation mode identifier, and unit time operation cost. The operation mode identifier includes general survey, detailed survey, and fixed-point operation. Based on these performance parameters, parametric models are created for each participating underwater vehicle, constructing parameter vectors for each vehicle. An underwater equipment cluster is then built based on these parameter vectors. :

[0078] ;

[0079] In the formula, Indicates the first Parameter vectors of underwater equipment in Taiwan This indicates the total number of underwater equipment involved in the search.

[0080] In S3, for each Assign an independent ant colony each include Each ant colony independently maintains two pheromone matrices: a path pheromone matrix and an allocation pheromone matrix. When an ant moves from a grid... Move to grid When constructing a grid To grid Path pheromone matrix , used to characterize Ants move their attraction from the current grid cell to build the grid. pheromone distribution matrix Used to characterize a grid Assigned to the Cumulative confidence level of Taiwan's underwater equipment.

[0081] In S4, the formula for calculating the state transition probability is:

[0082] ;

[0083] ;

[0084] ;

[0085] ;

[0086] In the formula, express time Ants from the current grid Select grid The probability is equivalent to time Ants in the grid cell The probability, express The ant path is a grid at any given time. To grid hour The path pheromone matrix is ​​equivalent to time Ants in the grid cell The path pheromone matrix is ​​used to record search efficiency and for search path planning. Indicates starting from the current grid cell Heuristic function Represents grid With grid The distance between them This indicates that the crashed target appeared in the grid. The probability of being hit is equivalent to the target crashing into a grid. The probability value, express time In the grid The pheromone distribution matrix is ​​used to characterize Time Grid Assigned to the The cumulative confidence level of Taiwan's underwater equipment. This indicates that the ant starts from the current grid cell. Select grid The turning cost factor For underwater equipment heading and grid Pointing grid The angle between directions, This indicates the included angle threshold, typically 60 degrees. This indicates that the ants are in the grid cells. Ocean current contributing factors, This indicates that the ants are in the grid cells. The angle between the course of underwater equipment and the direction of the ocean current is equivalent to an ant flying through a grid. Select grid At that time, the angle between the underwater equipment's heading and the ocean current direction, , , , , These represent the weight parameters corresponding to path pheromones, heuristic functions, allocation pheromones, turning cost factors, and ocean current assist factors, respectively. This represents the index of all candidate nodes, where each candidate node is the starting point of the ant's movement from the current grid. Selectable next grid, express The candidate node set includes all the next selectable grid cells for the ant from the current grid. express time Ants from the current grid Select candidate nodes Path pheromones, Indicates starting from the current grid to candidate node heuristic function, express time In candidate nodes The distribution of pheromones, This indicates that the ant starts from the current grid cell. Select candidate nodes The turning cost factor This indicates that the ant starts from the current grid cell. Select candidate nodes Ocean currents are a contributing factor.

[0087] , Obtained through iterative updates. The initial value is 0.1. The initial value is 0.01.

[0088] The ant calculates the state transition probability according to the formula, taking into account constraints such as target probability, equipment adaptability, turning cost, ocean current assistance, water depth obstacle, endurance, cable length, and safe distance, and selects the point with the highest state transition probability value as the next node.

[0089] The construction rules include:

[0090] (1) Candidate nodes are (2) The depth of the candidate node. Less than the Maximum diving depth of Taiwan's underwater equipment (3) Candidate nodes are non-obstacle-occupied grids in a heterogeneous grid diagram; (4) Current equipment remaining range In the formula Represents grid With grid The distance between them For the first (5) For ROV-type underwater equipment, the operating radius shall not exceed the cable length. (6) The distance between each underwater piece of equipment shall not be less than the dynamic safety distance. .

[0091] To provide detection probability, in constructing A certain number of repetitions are allowed, and the maximum number of times a grid can be repeated is set to replace candidate nodes. The construction rule that ants do not visit grids requires that, for ROV-type underwater equipment, the operating radius, i.e., the path length from the mother ship to the candidate node, must be less than the cable length. The terrain occlusion correction factor is a safety factor that takes into account terrain detours; it is typically set to 0.9 to meet the dynamic safety distance requirement. The construction rules include the shortest distance between the real-time candidate node and the planned path of other equipment. If the distance is less than the safety threshold, the candidate node will be excluded from the candidate node set to avoid equipment collision. The safety threshold is usually 200 meters.

[0092] In S5, an equipment grid depth adaptation constraint is introduced to construct the equipment grid depth adaptation matrix:

[0093] ;

[0094] In the formula, express Ants from the current grid To grid At that time, the first The compatibility between underwater equipment and grid selection in Taiwan. Represents a grid The water depth, For the first The optimal diving depth for Taiwan's underwater equipment. For the first The maximum diving depth of the underwater equipment is determined by embedding the constructed equipment grid depth fit matrix as a multiplicative factor into the state transition probability calculation formula as a heuristic function. .

[0095] In S6, based on the state transition rules after embedding equipment grid depth adaptation constraints, the ant colony algorithm is used to find the target search endpoint from the heterogeneous grid map. A meeting-based bidirectional path construction strategy is adopted, constructing paths from both the search start point and the target search endpoint simultaneously, and completing path splicing at the meeting point. The reference search path is output, and the fitness of ants in each independent ant colony is calculated. :

[0096] ;

[0097] In the formula, This represents the sum of the cumulative discovery probabilities of the path-covered raster. Represents the total probability over the entire domain. The latest time that underwater equipment can complete its search mission. Indicates the maximum allowed operation time for the search task. This indicates the total energy consumption of underwater equipment. This represents the expected energy consumption to complete the search task. This indicates the total depth of the operating grid beyond the maximum diving depth of the underwater equipment. This indicates the maximum allowable ultra-deep accumulation. , , , , These represent the weight coefficients of the corresponding items, and the sum of the five weight coefficients is 1.

[0098] During the construction of reference search paths, each independent ant colony updates its pheromones based on its fitness value, including local updates of path pheromones and global updates of pheromone allocation.

[0099] The local update of path pheromones is based on the local update rules of the ant colony algorithm. The path pheromones are updated immediately after each ant completes path construction, i.e., after an ant moves from one grid cell to the next.

[0100] ;

[0101] In the formula, express The ant path is a grid at any given time. To grid hour The path pheromone matrix represents the path pheromone matrix in the path pheromone matrix. Time, Grid To grid The pheromone concentration along this path, Indicates the local pheromone increment. Represents the locally updated coefficients. This indicates the pheromone volatile term. Indicates pheromone deposition term, The value range is from 0 to 1. The larger the pheromone level, the faster the historical pheromone decays, the shorter the ant's "memory" of the current path, and the stronger the exploratory nature of the ant colony algorithm. The smaller the value, the more historical pheromones are retained, resulting in stronger convergence of the ant colony algorithm.

[0102] By locally updating the path pheromones, each time an ant takes a step, some of the pheromones on the path will first evaporate, and then a fixed small amount will be added. When multiple ants frequently pass through the same path, the pheromones will slowly accumulate; conversely, the pheromones will gradually decay.

[0103] The global update of pheromone allocation follows a global update rule based on the MMAX (Maximum-Minimum Ant System) ant colony algorithm. After each independent ant colony constructs a complete reference search path, the fitness calculation formula is used to calculate... The fitness values ​​of each ant in the data are ranked first. Only a few ants receive an enhancement update, while the remaining ants receive a penalty decay update.

[0104] ;

[0105] ;

[0106] In the formula, express The ant path is a grid at any given time. To grid hour The pheromone distribution matrix represents the distribution of pheromones in ... Time, in the grid To grid This path is assigned pheromone concentration, Indicates the first All the grid cells visited by the ant Indicates the global update coefficient. Indicates the first The fitness of ants express The sum of the fitness of all ants is used to normalize the fitness of each ant.

[0107] The pheromone update is distributed globally according to the ant's fitness. Ants with higher fitness receive a greater weight in the pheromone increment along their path. The concept of a max-min ant system is introduced to... Limited to The interval, where, and As a fixed threshold artificially set based on engineering experience, and considering the application scenario of this invention and the initial pheromone value, the lower limit empirical value is usually taken. =0.01, maximum experience value =2.0; This represents the lower bound of the pheromone constraint, used to ensure that all paths have a certain probability of being selected, thus maintaining the exploratory nature of the ant colony algorithm. This represents the upper limit of pheromone limits, preventing the algorithm from stalling prematurely (early convergence) due to excessively high pheromone levels on a certain path, thus preventing the ant colony algorithm from losing its ability to explore other paths.

[0108] For various underwater equipment, especially AUVs with long endurance and large search area spans, a bidirectional ant colony strategy is adopted to accelerate the convergence of the ant colony algorithm. This involves using a bidirectional path construction strategy based on encounters to construct reference search paths, including... The ants are divided equally. and Two groups, The ants in the group start from the starting point of the reference search path (the search starting point of the underwater equipment, such as the mother ship deployment point) and search towards the boundary of the target hotspot area (the edge of the high-probability area identified by the probability map). The ants in a group start from the boundary of the target hotspot area or the target search endpoint in the heterogeneous grid (the point with the highest probability of the underwater landing location of the crashed target, determined by the ant colony algorithm), and search towards the starting point of the reference search path. When the search paths of two groups of ants intersect (i.e., there exists a node), the search proceeds. A group of ants passing through the grid ,and The ants in the group also passed through the grid. Record the connectivity of the intersecting nodes, trace back the search paths of the two groups of ants through the intersecting nodes, construct the complete path, and obtain the reference search path (i.e., the path between the two groups of ants). The group of ants moves from the starting point to the grid. The path, and The group of ants travels from the endpoint to the grid. The paths are concatenated to obtain a complete path from the starting point to the ending point. By employing a bidirectional path construction strategy, the initial blindness of unidirectional search is avoided, thus accelerating the convergence of the ant colony algorithm.

[0109] The ant colony algorithm iterates along the reference search path provided by S6 to simulate a search for a lost target, or performs an actual search for a lost target based on the reference search path provided by S6. After each iteration or each round of actual search, the probability of the underwater landing point of the lost target in the heterogeneous grid image is updated using Bayesian posterior analysis based on the iterative feedback or search feedback that no lost target was detected.

[0110] ;

[0111] In the formula, Indicates in During the iteration or search at any given moment, the crashed target is in the grid cell. The probability of its existence; Indicates in During the iteration or search at any given moment, the crashed target is in the grid cell. The probability of its existence; This indicates the probability of detection by the equipment.

[0112] A grid association rule-driven update strategy is introduced. When a suspected target feature is found in a grid, the probability of the neighboring grids of the grid containing the suspected feature is enhanced. The target occurrence probability of the grid and its neighbors (e.g., 3×3 or 5×5) is multiplied by an enhancement coefficient and then normalized. Simultaneously, this area is marked as a "suspicious area," which can guide more ants to verify it in subsequent iterations. The enhancement coefficient is greater than 1, and it is positively correlated with the confidence of the suspected feature; typically, it is set to 1.5.

[0113] When iterating over the reference search path provided by S6 using the ant colony algorithm, an iteration termination condition is set, including the maximum number of iterations. Iterative improvement of limit Among them, the maximum number of iterations A fixed threshold, set based on engineering experience, is used to balance the real-time requirements of underwater search missions with the algorithm's convergence needs, preventing the algorithm from getting stuck in an infinite loop. It is typically set to 300 to 500. When the algorithm iterations exceed the maximum number of iterations, or during the iteration process, continuous iterations... If the optimal solution does not improve, the iteration terminates and outputs the search area boundary of each underwater equipment, the optimal search path of each underwater equipment, and the expected cumulative discovery probability curve of the crashed target. Typically, a value of 50 is used. The search area boundary is described by polygons, representing the area boundary searched by each underwater vehicle. The optimal search path is represented by a 3D obstacle avoidance search path point set, where each underwater vehicle searches along its path. Path points include latitude, longitude, water depth, heading, and estimated time of arrival. The expected cumulative discovery probability curve for the crashed target represents the cumulative discovery probability over time and is used to assess the mission success rate.

[0114] An underwater target search path planning system based on ant colony algorithm, using the aforementioned underwater target search path planning method based on ant colony algorithm, includes a heterogeneous grid map construction module, a multi-ant colony cooperative control module, and a multi-equipment execution control module.

[0115] The heterogeneous grid map construction module includes multibeam echo sounder, side-scan sonar or synthetic aperture sonar, Doppler current profiler, satellite positioning system or beacon positioning system. It is used to obtain the probability map of the underwater landing point of the crashed target. In the obtained probability map of the underwater landing point of the crashed target, water depth and topography data, seabed obstacle distribution data and ocean current field data of the target sea area are integrated to construct a heterogeneous grid map to guide the search for the crashed target in the target sea area. Multibeam echo sounder is used to obtain water depth and topography data of the target sea area in real time, side-scan sonar or synthetic aperture sonar is used to obtain obstacle distribution in the target sea area, Doppler current profiler is used to obtain three-dimensional ocean current field data of the target sea area in real time, satellite positioning system is used to obtain the probability map of the underwater landing point of the crashed target through drift trajectory inversion, and beacon positioning system is used to obtain the probability map of the underwater landing point of the crashed target through acoustic beacon positioning.

[0116] The multi-ant colony collaborative control module includes an edge computing unit and an airborne computing unit, which are used to execute the ant colony algorithm to generate search path schemes. The edge computing unit is equipped with a three-layer heterogeneous pheromone matrix, including a path pheromone matrix, an allocation pheromone matrix, and an equipment grid depth adaptation matrix. The path pheromone matrix, the allocation pheromone matrix, and the heuristic function calculated based on the heterogeneous grid graph are coupled through a state transition probability calculation formula. The equipment grid depth adaptation matrix is ​​embedded in the state transition probability calculation formula as a multiplicative factor of the heuristic function. The airborne computing unit is deployed on the underwater equipment and is used to perform local path replanning and obstacle avoidance tasks.

[0117] The multi-equipment execution control module includes an equipment parameterization module, a path execution module, and an information feedback module. It guides the actual search tasks of each underwater device according to the search path plan and provides real-time feedback on the search results. The equipment parameterization module is used to perform parameterized modeling of the underwater devices participating in the search, obtain the parameter vectors corresponding to each underwater device, and build an underwater device cluster based on each parameter vector. The path execution module is used to generate search path plans according to the multi-ant colony collaborative control module and schedule each underwater device to participate in the actual search task of underwater crashed targets. The information feedback module is used to transmit the search results of crashed targets in real time. When the transmitted search result is that no target was detected, Bayesian posterior probability decay is performed. When the transmitted search result is that suspected target features or equipment failure was found, local ant colony replanning is triggered. The specific replanning mechanism includes (1) the impact on the landing probability of the crashed target: triggering the grid association rule, multiplying the probability of the target appearing in the grid where the suspected feature is located and its neighborhood by the enhancement coefficient to locally increase; (2) the impact on the pheromone matrix: increasing the pheromone concentration of the verification equipment (such as ROV) in the suspected area in the allocation pheromone matrix, or releasing the allocation pheromone of the area to which the faulty equipment belongs; (3) the impact on the state transition matrix: the above-mentioned local mutation of the target probability and pheromone directly updates the heuristic function and attraction weight in the state transition calculation formula, so that the state transition probability of the corresponding equipment converges sharply to the suspected area or the uncovered area, thereby realizing adaptive path change.

[0118] When constructing a heterogeneous raster map, the target sea area is discretized into a set of raster cells of the same size, with each raster cell having a size of [size missing]. Each grid stores attributes including the probability of a target being present in the grid, the grid's water depth, an obstacle occupancy flag in the grid, and the grid's ocean current vector. When the obstacle occupancy flag is 1, it indicates that the grid is occupied by an obstacle and underwater equipment cannot pass through.

[0119] The ant colony algorithm of the multi-ant colony collaborative control module is a multi-ant colony collaborative planning algorithm based on a multi-ant colony and multi-equipment collaborative architecture. It is used to perform multi-ant colony collaborative search for search paths. This includes assigning an independent ant colony to each parameter vector in the underwater equipment cluster. Each ant colony maintains a path pheromone matrix and an allocation pheromone matrix. By constructing an ant colony-equipment mapping mechanism and introducing equipment grid depth adaptation constraints, path planning is performed, thereby realizing the joint solution of region segmentation, equipment allocation, and path planning.

[0120] At least two types of underwater equipment should be selected from the following: intelligent underwater robot (AUV), remotely operated underwater robot (ROV), and manned underwater vehicle (HOV). Performance parameters of each piece of underwater equipment participating in the search should be read, including maximum diving depth, endurance, cruising speed, swath width, turning radius, operation mode identifier, and unit time operation cost. The operation mode identifier includes general survey, detailed survey, and fixed-point operation. Based on the performance parameters of each piece of underwater equipment, parametric modeling should be performed on each piece of equipment participating in the search, constructing a parameter vector for each piece of equipment. An underwater equipment cluster should then be constructed based on the parameter vectors of each piece of equipment. :

[0121] ;

[0122] In the formula, Indicates the first Parameter vectors of underwater equipment in Taiwan This indicates the total number of underwater equipment involved in the search.

[0123] equipment The parameter vector includes the following parameters:

[0124] Maximum diving depth Unit is meters, equipment endurance The unit is hour, and the equipment's cruising speed is... The unit is a knot (kn), and the equipment's detection swath width is... The unit is meters, and the minimum turning radius of the equipment is... Units are meters, and the cost per unit of time is [not specified]. Unit is yuan per hour, equipment operation mode identifier. , This indicates that the equipment is undergoing a general survey. The time indicates that a detailed inspection of the equipment is underway. This indicates that the equipment is performing a fixed-point operation task.

[0125] For AUV-type underwater equipment, it is used to guide the execution of large-area gridded probabilistic survey tasks; for ROV-type underwater equipment, it is used to guide the execution of detailed verification and sample collection tasks in suspicious areas (detailed investigation tasks); for HOV-type underwater equipment, it is used to guide the execution of manned verification and decision confirmation tasks in core high-probability areas (fixed-point operation tasks).

[0126] Based on a multi-ant colony, multi-device collaborative architecture, for each Assign an independent ant colony Each include Just an ant. express The population size is typically between 30 and 100.

[0127] Each independent population maintains path pheromones and allocation pheromones. Path pheromones characterize the attractiveness of a path to ants, while allocation pheromones characterize the cumulative confidence level of path allocation to each equipment. The initial value is Typically, the value is set to 0.1, and the initial value for pheromone allocation is... The multi-equipment execution control module is also equipped with an underwater communication network consisting of an underwater acoustic communication modem, underwater WiFi, and underwater optical communication, which is used to transmit the position and detection data of the underwater equipment in real time.

[0128] Example 1: Search for the black box of a crashed aircraft in a certain sea area;

[0129] Mission Scenario: A civilian airliner has crashed into the sea, and the search and rescue operation is underway to locate the black box equipped with a 37.5kHz beacon. The search and rescue center has deployed 4 AUVs (Explorer 1000 type), 2 ROVs (Sea Lion type), and 1 HOV (Deep Sea Warrior), with the mother ship being the "Explorer One". The mission area is 20 nautical miles x 20 nautical miles, with a water depth of 200-1200 meters, and seamounts in the northwest.

[0130] Parameter settings: Grid size 100m × 100m, ant colony size Weight .

[0131] Implementation process: Follow steps 2.1 to 2.10. After 230 iterations, the solution converges, and the output is as follows:

[0132] AUV01 covers the northwest shallow water area (25km²), AUV02 covers the southeast deep-sea plain (30km²), and AUV03 / 04 cover the central transition zone respectively.

[0133] ROV01 / 02 are responsible for detailed verification of high-probability areas, while HOV01 is responsible for manual verification of core areas.

[0134] The expected cumulative detection probability over 6 hours is 79.8%.

[0135] Comparative experiments: Compared with the zigzag sea sweep (manual partitioning), the cumulative discovery probability increased by 94%; compared with the standard ant colony algorithm (without pheromone allocation), it increased by 18.6%, and the convergence speed increased by 42%. The equipment achieved zero violations in ultra-deep areas, and the coverage rate of high-probability areas reached 94%.

[0136] Example 2: Underwater archaeological survey of a shipwreck site;

[0137] Mission Scenario: Locating a Song Dynasty shipwreck site. The known area is 5 nautical miles x 5 nautical miles, with a water depth of 50–80 meters. Equipment: An AUV equipped with a magnetometer (50-meter detection width), an ROV equipped with a hyperspectral imager, and a HOV for expert interpretation.

[0138] Adaptive adjustments: A magnetic detection fitness factor is added to the fitness matrix, and an archaeological value gradient is introduced into the heuristic function (the probability of the neighborhood increases after the discovery of porcelain shards). Dynamic updates use the "shipwreck debris" association rule to generate an elliptical probability enhancement zone centered on the discovery point.

[0139] Results: The main body of the sunken ship was located within 8 hours, which is 3 times more efficient than traditional grid search.

[0140] Example 3: Emergency search and rescue of a deep-sea submarine in distress;

[0141] Mission Scenario: The wrecked submarine is located in a complex seamount terrain area, with a water depth of 800-3000m and a golden rescue time of 72 hours. Equipment: 2 long-range AUVs (endurance 40h, detection swath width 500m), 1 deep-sea ROV (4500m), and 1 HOV.

[0142] Key implementation: Enhance the penalty for terrain ruggedness during state transitions.

[0143] ROV operating radius dynamic constraint: calculate the terrain detour path length from the mother ship to the target point in real time to ensure it is less than the cable length.

[0144] Emergency redistribution mechanism: When an AUV discovers suspected debris, its general survey mission is suspended and switched to detailed scanning. The ROV is notified to conduct a verification, and the paths of other equipment are updated to avoid conflicts.

[0145] Fitness weight adjustment: Increased to 0.4, Increased to 0.2.

[0146] Results: Global planning was completed within 1 hour, and the AUV detected an abnormal heat source within 3 hours, guiding the ROV to confirm the location of the wrecked submarine, thus buying time for the rescue.

[0147] To verify the effectiveness of the underwater target search path planning method and system based on ant colony algorithm provided by this invention, simulation experiments were conducted in typical sea areas. The effectiveness of this invention was compared with that of standard ant colony algorithm and traditional scanning method. The experimental results are shown in Table 1.

[0148] Table 1 Comparison of simulation experiment results;

[0149] .

[0150] As can be seen from Table 1, compared with the standard ant colony algorithm and traditional scanning methods, the path planning method provided by this invention significantly improves search efficiency, convergence speed, resource adaptability, and dynamic response capability.

[0151] Figure 1This is a probability distribution map of the initial impact point of the crashed target. Red areas represent high-probability density areas, blue areas represent low-probability density areas, and green circles represent areas within the lower-probability density that are likely to be the impact points of the crashed target. Search equipment allocation and search path planning are based on this initial impact point probability map.

[0152] Figure 2 Assign maps to search devices, Figure 1 The search area is divided into 20 sub-regions. Based on the probability of landing in each sub-region, the search areas are further categorized into three types: critical intervention, detailed investigation, and general survey. The required search equipment is determined according to the type of search area.

[0153] Figure 3 The search equipment path planning map includes search paths for AUVs, ROVs, and HOVs. For AUVs, it guides large-area grid-based probabilistic surveys; for ROVs, it guides detailed verification and sample collection in suspicious areas (detailed investigation); and for HOVs, it guides manned verification and decision confirmation in high-probability areas (fixed-point operation). Search areas are divided based on landing probability, determining the search equipment and corresponding search paths. When the landing probability is greater than 50%, the search area is considered critical for intervention, and HOVs are prioritized, with a spiral search path. If HOVs are insufficient or for other reasons, ROVs are used secondarily, with a zigzag search path. When the landing probability is between 5% and 50%, the search area is considered for detailed investigation, and ROVs are prioritized; when the landing probability is less than 5%, the search area is considered a general survey, and AUVs are prioritized, with a lawnmower search path.

[0154] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for underwater target search path planning based on ant colony algorithm, characterized in that, include: S1. Construct a heterogeneous grid map, including obtaining a probability map of the underwater impact point of the crashed target, and integrating water depth topography data, seabed obstacle distribution data, and ocean current field data; S2. Parametrically model each underwater equipment involved in the search and construct an underwater equipment cluster; S3. Construct an ant colony equipment mapping mechanism, assign an independent ant colony to each underwater equipment, and each independent ant colony maintains the path pheromone and allocation pheromone of the corresponding equipment. S4. Construct state transition rules and calculate the state transition probabilities of coupled path pheromones, allocated pheromones, and heterogeneous grid diagrams; S5. Introduce equipment grid depth adaptation constraints to eliminate grids that underwater equipment cannot reach in heterogeneous grid maps; S6. Adopting a bidirectional path construction strategy based on encounters, a reference search path is constructed based on constraint state transition rules and equipment grid depth adaptation constraints. S7. Based on the reference search path provided by S6, search for the crashed target. During the search process, according to the search feedback of each underwater equipment, update the heterogeneous grid map of S1 through Bayesian posterior and output the search area boundary and optimal search path of each underwater equipment. In S1, heterogeneous raster diagram for: ； In the formula, This represents the probability matrix indicating the underwater impact point of the crashed target. This represents the water depth matrix of the target sea area. This represents the obstacle occupancy matrix in the target area. This represents the ocean current intensity matrix for the target sea area, which is the area where the lost target is to be salvaged. In S2, underwater equipment cluster for: ； In the formula, Indicates the first Parameter vectors of underwater equipment in Taiwan This indicates the total number of underwater equipment involved in the search; During the construction of reference search paths, each independent ant colony updates its pheromones based on its fitness value, including local updates of path pheromones and global updates of pheromone allocation. The local update of path pheromones is based on the local update rules of the ant colony algorithm, whereby the path pheromones are updated after each ant completes its path construction: ； In the formula, express time Ants in the grid cell The path pheromone matrix, representing the path pheromone matrix in Time grid unit pheromone concentration, Indicates the local pheromone increment. Represents the local update coefficients. This indicates the pheromone volatile term. This indicates the pheromone deposition term.

2. The underwater target search path planning method based on ant colony algorithm according to claim 1, characterized in that, In S3, for each Assign an independent ant colony Each include Each ant colony independently maintains two pheromone matrices, including a path pheromone matrix and an allocation pheromone matrix. In S4, the state transition rules, coupling path pheromone matrix, allocation pheromone matrix, and heterogeneous grid map are used to calculate the state transition probability using the following formulas: ； ； ； ； In the formula, express time Ants in the grid cell The probability, express time Ants in the grid cell The path pheromone matrix, Represents grid cells heuristic function, Represents a grid With grid The distance between them This indicates that the crashed target appeared in the grid. The probability in express time In the grid The allocation of pheromone matrix, This indicates that the ants are in the grid cells. The turning cost factor For underwater equipment heading and grid Pointing grid The angle between directions, Indicates the included angle threshold. This indicates that the ants are in the grid cells. Ocean current contributing factors, This indicates that the ants are in the grid cells. The angle between the course of the underwater equipment and the direction of the ocean current. , , , , These represent the weight parameters corresponding to path pheromones, heuristic functions, allocation pheromones, turning cost factors, and ocean current assist factors, respectively. Indicates the index of the candidate node. express The candidate node set, express time In grid cells Path pheromones, Represents grid cells heuristic function, express time In candidate nodes The distribution of pheromones, This indicates that the ants are in the grid cells. The turning cost factor This indicates that the ants are in the grid cells. Ocean currents are a contributing factor.

3. The underwater target search path planning method based on ant colony algorithm according to claim 2, characterized in that, In S5, an equipment grid depth adaptation constraint is introduced to construct the equipment grid depth adaptation matrix: ； In the formula, Indicates the first Taiwan Underwater Equipment and Grid Unit Adaptability Represents a grid The water depth, For the first The optimal diving depth for Taiwan's underwater equipment. For the first The maximum diving depth of Taiwan's underwater equipment For the first The minimum diving depth of the underwater equipment is determined by embedding the constructed equipment grid depth adaptation matrix as a multiplicative factor into the state transition probability calculation formula as a heuristic function.

4. The underwater target search path planning method based on ant colony algorithm according to claim 3, characterized in that, In S6, based on the state transition rules after embedding equipment grid depth adaptation constraints, the ant colony algorithm is used to find the target search endpoint from the heterogeneous grid map. A meeting-based bidirectional path construction strategy is adopted, constructing paths from both the search start point and the target search endpoint simultaneously, and completing path splicing at the meeting point. The reference search path is output, and the fitness of ants in each independent ant colony is calculated. : ； In the formula, This represents the sum of the cumulative discovery probabilities of the path-covered raster. Represents the total probability over the entire domain. The latest time that underwater equipment can complete its search mission. Indicates the maximum allowed operation time for the search task. This indicates the total energy consumption of underwater equipment. This represents the expected energy consumption to complete the search task. This indicates the total depth of the operating grid beyond the maximum diving depth of the underwater equipment. This indicates the maximum allowable ultra-deep accumulation. , , , , These represent the weight coefficients of the corresponding items.

5. The underwater target search path planning method based on ant colony algorithm according to claim 4, characterized in that, The global update of pheromone allocation follows a global update rule based on the MMAX (Maximum-Minimum) ant colony algorithm. After each independent ant colony constructs a complete reference search path, the fitness calculation formula is used to calculate... The fitness values ​​of each ant in the data are ranked from the top to the bottom. Only a few ants receive an enhancement update, while the remaining ants receive a penalty decay update. ； ； In the formula, express time Ants in the grid cell The allocation of pheromone matrix, Indicates the first All the grid cells visited by the ant Indicates the global update coefficient. Indicates the first The fitness of ants express The sum of the fitness of all ants is used to normalize the fitness of each ant.

6. The underwater target search path planning method based on ant colony algorithm according to claim 5, characterized in that, The ant colony algorithm is used to search for the crashed target on the reference search path provided by S6. During the search process, based on the search feedback that no crashed target was detected, the probability of the underwater landing point of the crashed target in the heterogeneous grid image is updated using Bayesian posterior. ； In the formula, Indicates in During the ongoing search, the crashed target was located within a grid cell. The probability of; Indicates in During the ongoing search, the crashed target was located within a grid cell. The probability of; Indicates the probability of equipment detection; Set iteration termination conditions, including the maximum number of iterations. Iterative improvement of limit When the algorithm iterations exceed the maximum number of iterations, or during the iteration process, continuous iterations... If the optimal solution does not improve, the iteration terminates and outputs the search area boundary of each underwater equipment, the optimal search path of each underwater equipment, and the expected cumulative discovery probability curve of the crashed target.

7. An underwater target search path planning system based on ant colony algorithm, characterized in that, The underwater target search path planning method based on ant colony algorithm as described in claim 1 includes a heterogeneous grid graph construction module, a multi-ant colony collaborative control module, and a multi-equipment execution control module. The heterogeneous grid map construction module includes multibeam echo sounder, side-scan sonar or synthetic aperture sonar, Doppler current profiler, satellite positioning system or beacon positioning system, used to obtain probability maps of the underwater landing points of the crashed target, and to integrate water depth and topography data, seabed obstacle distribution data and ocean current field data of the target sea area into the obtained probability maps of the underwater landing points of the crashed target, in order to construct a heterogeneous grid map to guide the search for the crashed target in the target sea area; The multi-ant colony collaborative control module includes an edge computing unit and an airborne computing unit, which are used to execute the ant colony algorithm to generate search path schemes. The edge computing unit is equipped with a three-layer heterogeneous pheromone matrix, including a path pheromone matrix, an allocation pheromone matrix, and an equipment grid depth adaptation matrix. The path pheromone matrix, the allocation pheromone matrix, and the heuristic function calculated based on the heterogeneous grid graph are coupled through a state transition probability calculation formula. The equipment grid depth adaptation matrix is ​​embedded in the state transition probability calculation formula as a multiplicative factor of the heuristic function. The airborne computing unit is deployed on the underwater equipment and is used to perform local path replanning and obstacle avoidance tasks. The multi-equipment execution control module includes an equipment parameterization module, a path execution module, and an information feedback module. It is used to guide the actual search tasks of each underwater equipment according to the search path plan and to provide real-time feedback on the search results.

8. The underwater target search path planning system based on ant colony algorithm according to claim 7, characterized in that, Multibeam echo sounders are used to acquire real-time water depth and topographic data of the target sea area; side-scan sonar or synthetic aperture sonar is used to acquire the distribution of obstacles in the target sea area; Doppler current profilers are used to acquire real-time three-dimensional ocean current field data of the target sea area; satellite positioning systems are used to obtain the probability map of the underwater landing point of the crashed target through drift trajectory inversion; and beacon positioning systems are used to obtain the probability map of the underwater landing point of the crashed target through acoustic beacon positioning. The equipment parameterization module is used to perform parameterized modeling of the underwater equipment participating in the search, obtain the parameter vectors corresponding to each underwater equipment, and build an underwater equipment cluster based on each parameter vector. The path execution module is used to generate search path schemes according to the multi-ant colony collaborative control module, and schedule each underwater equipment to participate in the actual search task of underwater crashed targets. The information feedback module is used to transmit the search results of crashed targets in real time. When the transmitted search result is that no target was detected, Bayesian posterior probability decay is performed. When the transmitted search result is that suspected target features or equipment failure is found, local ant colony replanning is triggered.

Images