Global path planning method for robot inspection of smart wind farm booster station
By pre-setting empirical parameters and dynamically adjusting ant colony algorithm parameters in the robot inspection of wind farm booster stations, the path planning process is simplified, solving the problems of high computational resource consumption and insufficient real-time performance in existing technologies, and realizing efficient inspection with rapid response to complex environments.
Patent Information
- Application Number
- CN202410590236.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-13
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2044-05-13
AI Technical Summary
Existing global path planning technology for robot inspection of wind farm booster stations, which combines flower pollination algorithm and improved ant colony algorithm, results in long computation time and high resource consumption, making it difficult to meet the real-time requirements and inspection applications with limited hardware resources.
By adopting preset empirical parameter values (α0, β0, ρ0) and a dynamic adjustment strategy based on historical optimal solutions, the parameter optimization process of the ant colony algorithm is simplified, and the pheromone importance factor, heuristic function importance factor, and volatile factor are dynamically adjusted to optimize path planning.
It significantly reduces computation time and resource consumption, improves the real-time performance and adaptability of path planning, and is suitable for wind farm booster station inspection tasks with limited hardware resources or high real-time requirements.
Smart Images

Figure CN118348997B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of resource allocation, specifically relating to a global path planning method for robot inspection of smart wind farm booster stations. Background Technology
[0002] Existing global path planning technology for robotic inspection of wind farm booster stations combines the Flower Pollination Algorithm (FPA) with an improved ant colony algorithm to address inspection needs in complex operating environments. The main steps include:
[0003] 1. Parameter optimization: The parameters (pheromone importance factor α, heuristic function importance factor β, and volatile factor ρ) of the ant colony algorithm are optimized using FPA to simulate the pollination process of flowers (self-pollination and cross-pollination). The feasible solution is updated iteratively until the termination condition is met, and the optimal parameters are output.
[0004] 2. Parameter assignment: Set the ant colony algorithm parameters, including the total number of ants m, the iteration threshold N, and α, β, and ρ obtained from FPA; 3. Improve ant colony iteration: Place m ants randomly on nodes, select the direction of movement based on pheromone concentration and expectation, traverse all nodes and return to the starting point to form a complete loop; calculate the cost of each loop and select the one with the lowest cost as the current optimal path;
[0005] 4. Path Update: After completing one iteration, update the pheromone concentration of each branch in the optimal path set, and adjust the accumulation and volatilization of pheromones with each iteration according to the formula;
[0006] 5. Branch contribution calculation: Calculate the contribution ratio of each branch in the optimal path, and add extra pheromones to branches that exceed the threshold;
[0007] 6. Iteration Judgment: Check if the number of iterations has reached the threshold. If yes, output the globally optimal inspection path; otherwise, return to the parameter optimization step to continue optimization.
[0008] Because this existing technology combines the Flower Pollination Algorithm (FPA) with an improved ant colony algorithm, it involves multiple optimization steps and parameter adjustments, making the entire path planning process relatively complex. The complex algorithm structure leads to long computation time, especially in scenarios with a large number of nodes, such as wind farm booster stations, where multiple iterations and multi-step optimizations consume a lot of computing resources. This is not suitable for inspection applications with high real-time requirements or limited hardware resources. Summary of the Invention
[0009] The purpose of this invention is to provide a global path planning method for the inspection of intelligent wind farm booster stations by robots, so as to solve the problems mentioned in the background art.
[0010] To solve the above-mentioned technical problems, the present invention provides the following technical solution:
[0011] A global path planning method for robot inspection of smart wind farm booster station includes the following steps: Step S0: Preset a set of empirical parameter values for improving the ant colony algorithm;
[0012] Step S1: Initialize the improved ant colony algorithm, setting the total number of ants m and the iteration threshold N;
[0013] Step S2: Perform the ant colony algorithm iterative process; at the beginning of each iteration, dynamically adjust the pheromone importance factor α, the heuristic function importance factor β, and the volatile factor ρ based on the historical evolution of the current global optimal solution;
[0014] Step S3 uses an improved ant colony algorithm to iterate, select the path with the lowest cost, update the pheromone concentration, calculate the contribution ratio of branches, and determine the termination condition of the iteration until the iteration number threshold is reached or other termination conditions are met.
[0015] Furthermore, the empirical parameter values for step S0 specifically include:
[0016] Pheromonic importance factor α: set to the empirical value α0;
[0017] The heuristic function importance factor β: set to the empirical value β0;
[0018] Volatility factor ρ: Set to the empirical value ρ0;
[0019] Empirical parameter values are selected based on the environmental characteristics of the wind farm's booster station and the requirements of the inspection tasks.
[0020] Furthermore, step S1 specifically involves:
[0021] Set the total number of ants to `m`;
[0022] Set a maximum iteration threshold `N` as one of the conditions for the algorithm to stop searching. If no better solution is found after exceeding this number of iterations, the algorithm is considered to be close to the optimal solution or trapped in a local optimum.
[0023] Construct a network topology diagram of the wind farm booster station, where each node represents the equipment or area to be inspected, and each edge represents the actual path connecting two nodes, with the corresponding distance or cost attached.
[0024] Initialize the pheromone matrix, ensuring a uniform initial pheromone concentration distribution across all paths or assigning certain prior information based on historical inspection data; initialize the global optimal solution record, including the current shortest path and its corresponding cost;
[0025] Each ant is randomly placed on a starting node and given an empty list of paths, ready to begin traversing the network to find the optimal path.
[0026] Further, step S2 includes: step S2.1: calculate the interval Δt between the current iteration number t and the most recent global optimal solution update iteration number t0; calculate the current iteration number `t`; obtain and record the iteration number `t0` when the global optimal solution was last found;
[0027] The time interval Δt = t - t0 is calculated based on the current iteration number `t` and the iteration number `t0` updated by the most recent global optimal solution; Δt reflects whether the current algorithm is trapped in a local optimum or the degree of effective search in the global scope.
[0028] Furthermore, step S2 includes: step S2.2: adjusting α, β, and ρ according to Δt and preset adjustment rules;
[0029] Step S2.2 in detail:
[0030] If Δ`t` is large and greater than the threshold, increase the value of the volatile factor ρ to accelerate the dissipation of pheromones, prompting ants to explore more unconsidered paths and break the existing local optimum trap; if Δ`t` is small and less than the threshold, maintain the volatile factor ρ or decrease the volatile factor ρ to maintain the current good search space convergence performance.
[0031] Adjust α and β based on Δ`t`, the quality change of the local solution, and the remaining number of iterations.
[0032] Furthermore, step S2 includes: Step S2.3: Updating the adjusted α, β, and ρ values for path selection by ants within the current iteration cycle; Step S2.3 specifically:
[0033] Set specific parameter adjustment rules, including linear increase and exponential decay, and calculate new α', β', and ρ' values based on the magnitude of Δ`t`;
[0034] The adjusted new parameter values α', β', and ρ' are applied to the ant path selection process in the current iteration cycle;
[0035] After each iteration cycle, check the algorithm performance to see if the new parameter settings have improved the search performance;
[0036] The updated parameters are used to continue the iterative process of the ant colony algorithm, including path selection, pheromone update, and branch contribution ratio calculation, until the preset iteration number threshold N is reached.
[0037] Furthermore, step S3 is detailed as follows:
[0038] Within the current iteration period t, each virtual ant selects a path according to the improved ant colony algorithm; each ant determines the probability of moving between nodes in the graph based on the adjusted parameter values α, β, and ρ; specifically, the probability of each node being selected as the next visit point consists of two parts: the selection probability based on pheromone intensity and the selection probability based on heuristic information.
[0039] All ants traverse the entire map network of the booster station, construct their own inspection paths, and calculate the corresponding path costs. After completing one round of iteration, the pheromone matrix is updated based on the path costs found by the ants. For the edges traversed by ants that have discovered better paths, the corresponding pheromone amount is increased. The pheromone on all edges is decayed according to the evaporation factor ρ.
[0040] Compare the costs of all paths found by all ants. If an ant finds a globally optimal path with a lower cost, update the globally optimal solution and record the iteration number t0 of this update.
[0041] Based on the importance and contribution of each branch in the optimal path, the weighting coefficient of its pheromone update is further calculated to guide the subsequent ants' search on key branches;
[0042] Check if the preset iteration count threshold N has been reached; if the threshold has not been reached, continue to the next iteration; if the threshold has been reached but the global optimal solution still does not meet the preset convergence criterion, extend the iteration count.
[0043] In each iteration, the three core parameters α, β, and ρ are continuously monitored and adjusted according to the dynamic adjustment strategy in step S2.2 to adapt to environmental changes and task requirements, enabling the algorithm to explore the global optimal path more effectively at different stages.
[0044] Furthermore, path costs include, but are not limited to, path length, expected inspection time, and costs of avoiding potentially hazardous areas.
[0045] Beneficial effects: This application eliminates the parameter optimization process of the flower pollination algorithm by directly setting empirical parameter values, thus reducing the optimization level. Simultaneously, it replaces the complex pollination mechanism with a dynamic adjustment strategy based on historical optimal solutions, maintaining a certain degree of adaptability while simplifying the algorithm structure. This significantly reduces computation time and required computational resources, improving the real-time performance of the method and making it more suitable for use in wind farm booster station inspection applications with limited hardware resources or high real-time requirements. In practical applications, this method is expected to achieve rapid response to complex operating environments while ensuring inspection efficiency and effectiveness. Attached Figure Description
[0046] Figure 1A flowchart illustrating the global path planning method for robot inspection of smart wind farm booster stations. Detailed Implementation
[0047] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0048] This application discloses a global path planning method for robot inspection of smart wind farm booster stations, referencing... Figure 1 The process includes steps, such as setting empirical parameter values. Step S0: Based on practical experience or expert recommendations, a set of empirical parameter values is preset to improve the ant colony algorithm. Step S0 specifically includes:
[0049] Pheromone importance factor α: set to an empirical value of α0, reflecting the weight of pheromones in guiding ant decision-making;
[0050] Heuristic function importance factor β: set to an empirical value of β0, reflecting the role of heuristic information in path selection;
[0051] Volatilization factor ρ: set to an empirical value ρ0 to control the rate at which pheromones decay with the number of iterations;
[0052] Empirical parameter values are selected based on the environmental characteristics of the wind farm booster station and the requirements of the inspection task to ensure that relatively ideal path planning results can be obtained in most scenarios;
[0053] A simplified pollination machine, based on a dynamic adjustment strategy using historical optimal solutions.
[0054] Step S1: Initialize the improved ant colony algorithm, set the total number of ants m, the iteration threshold N, and other fixed parameters; cancel the use of the flower pollination algorithm and directly use the preset empirical parameter values;
[0055] Step S1 is as follows:
[0056] Set the total number of ants `m`, which is a key parameter that affects the algorithm's global search capability. It represents the number of virtual ants participating in the search at the same time and is determined based on the specific scale and complexity of the wind farm's booster station and the desired search coverage.
[0057] Set a maximum iteration threshold `N` as one of the conditions for the algorithm to stop searching. If no better solution is found after exceeding this number of iterations, the algorithm is considered to be close to the optimal solution or trapped in a local optimum.
[0058] Construct a network topology diagram of the wind farm booster station, where each node represents the equipment or area to be inspected, and each edge represents the actual path connecting two nodes, with the corresponding distance or cost attached.
[0059] Initialize the pheromone matrix, ensuring a uniform initial pheromone concentration distribution across all paths or assigning certain prior information based on historical inspection data; initialize the global optimal solution record, including the current shortest path and its corresponding cost;
[0060] Each ant is randomly placed on a starting node and given an empty list of paths, ready to begin traversing the network to find the optimal path.
[0061] Consider setting the ant movement probability transfer rules according to the actual situation, such as based on pheromone concentration, heuristic function (such as the actual distance between nodes) and other factors that affect path selection;
[0062] Step S2: Perform the ant colony algorithm iterative process; at the beginning of each iteration, dynamically adjust the pheromone importance factor α, the heuristic function importance factor β, and the volatile factor ρ based on the historical evolution of the current global optimal solution;
[0063] Step S2 includes:
[0064] Step S2.1: Calculate the interval Δt(t-t0) between the current iteration number t and the iteration number t0 of the most recent global optimal solution update;
[0065] Step S2.1 Specifically, calculate the current iteration number `t`; obtain and record the iteration number `t0` when the global optimal solution was last found;
[0066] The time interval Δt is calculated based on the current iteration number `t` and the iteration number `t0` updated by the most recent global optimal solution (`t0`). Δt reflects whether the current algorithm is trapped in a local optimum or the degree of effective search in the global scope.
[0067] Step S2.2: Fine-tune α, β, and ρ according to Δt and the preset adjustment rules;
[0068] Step S2.2 in detail:
[0069] If Δ`t` is large, it means that no better path has been found for a long time, which means that the algorithm has problems with premature convergence or getting trapped in local optima. In this case, the value of the volatile factor ρ should be increased to accelerate the dissipation of pheromones and encourage ants to explore more unconsidered paths and break the existing local optimum trap. If Δ`t` is small, the algorithm is still effectively finding better solutions. In this case, the volatile factor ρ should be kept at a low level, or even slightly reduced, to maintain the current good search space convergence performance.
[0070] Based on Δ`t` and other performance indicators (such as changes in the quality of local solutions, remaining iterations, etc.), α and β are adjusted accordingly to balance the algorithm's global exploration capability and local development capability.
[0071] When it is necessary to enhance exploration, increase α to emphasize the pheromone accumulation of previously taken excellent paths, or adjust β appropriately to reduce the influence of the heuristic function and encourage the attempt of different path combinations;
[0072] Conversely, if it is necessary to enhance the local search, α can be adjusted to reduce the influence of pheromones, β can be increased to rely on heuristic information, and paths with lower expected costs can be prioritized.
[0073] Step S2.3: Update the adjusted α, β, and ρ values for path selection by ants in the current iteration cycle;
[0074] Step S2.3 in detail:
[0075] Set specific parameter adjustment rules, such as linear increment, exponential decay or other adaptive strategies, and calculate new α', β', and ρ' values based on the magnitude of Δ`t`;
[0076] The adjusted new parameter values α', β', and ρ' are applied to the ant path selection process in the current iteration cycle;
[0077] After each iteration cycle, check the algorithm performance to see if the new parameter settings have improved the search performance. If necessary, continue to dynamically adjust in subsequent iterations.
[0078] The ant colony algorithm continues to be executed using the updated parameters, including steps such as path selection, pheromone update, and branch contribution ratio calculation, until the preset iteration number threshold N is reached, or other early termination conditions are met (e.g., when there is no significant improvement after several consecutive iterations).
[0079] Step S3 uses an improved ant colony algorithm to iterate, select the path with the lowest cost, update the pheromone concentration, calculate the branch contribution ratio, and determine the iteration termination condition until the iteration number threshold is reached or other termination conditions are met.
[0080] Step S3 is detailed as follows:
[0081] Within the current iteration period t, each virtual ant selects a path according to the improved ant colony algorithm; each ant determines the probability of moving between nodes in the graph based on the adjusted parameter values α, β, and ρ; specifically, the probability of each node being selected as the next visit point consists of two parts: the selection probability based on pheromone intensity and the selection probability based on heuristic information (such as distance, estimated remaining energy, task priority, etc.).
[0082] All ants traverse the entire map network of the booster station, construct their own inspection paths, and calculate the corresponding path costs; the costs typically consider factors including but not limited to path length, expected inspection time, and the cost of avoiding potentially dangerous areas.
[0083] After completing one iteration, the pheromone matrix is updated based on the path cost found by the ants; for the edges traversed by the ants that have found a better path, the corresponding pheromone amount is increased; and the pheromone on all edges is decayed to a certain extent according to the evaporation factor ρ, in order to balance exploitation (using known information) and exploration (exploring unknown areas).
[0084] Compare the costs of all paths found by all ants. If an ant finds a globally optimal path with a lower cost, update the globally optimal solution and record the iteration number t0 of this update.
[0085] Based on the importance and contribution of each branch in the optimal path, the weighting coefficient of its pheromone update can be further calculated, which helps to more accurately guide the subsequent ants' search on key branches.
[0086] Check if the preset iteration threshold N has been reached; if the threshold has not been reached, continue to the next round of iteration; if the threshold has been reached but the global optimal solution still does not meet the preset convergence criteria or other termination conditions (such as no significant improvement after several consecutive iterations), the number of iterations can be appropriately extended or other strategies can be adopted to try to escape the local optimum.
[0087] In each iteration, the three core parameters α, β, and ρ are continuously monitored and adjusted according to the dynamic adjustment strategy described in step S2.2 to adapt to environmental changes and task requirements, enabling the algorithm to explore the global optimal path more effectively at different stages.
[0088] Compared with the existing technologies mentioned in the background, the global path planning method for robot inspection of smart wind farm booster stations proposed in this application has the following significant technical advantages:
[0089] 1. Simplified optimization process: This application no longer relies on the flower pollination algorithm (FPA) to optimize the ant colony algorithm parameters, but directly presets a set of empirical parameter values (α0, β0, ρ0), which greatly simplifies the optimization level and avoids the additional computational overhead and complexity brought by FPA.
[0090] 2. Dynamic adjustment strategy: This application innovatively adopts the core parameters (α, β, ρ) of the ant colony algorithm to be dynamically adjusted based on the historical best solution. These parameters are adjusted in a timely manner according to the time interval Δt of the current global best solution update iteration, which can more flexibly respond to environmental changes and inspection task requirements, enhance the algorithm's adaptability, and simplify the original pollination mechanism.
[0091] 3. Improved computational efficiency: By abandoning FPA parameter optimization and complex pollination mechanisms, the method in this application significantly reduces computation time and required computational resources, improving the real-time performance of global path planning. It is particularly suitable for wind farm booster station inspection application scenarios with limited hardware resources or high real-time requirements.
[0092] 4. Enhanced practicality: While ensuring inspection efficiency and effectiveness, the method of this application can better cope with inspection tasks such as wind farm substations with a large number of nodes and complex environments. Through dynamically adjusted parameter strategies, the robot can quickly respond to environmental changes and achieve efficient inspection of complex operating environments.
[0093] In summary, the global path planning method for the inspection of the smart wind farm booster station robot provided in this application, while retaining the advantages of the original ant colony algorithm, effectively solves the problems of high computational resource consumption and insufficient real-time performance in the existing technology by simplifying the optimization process and introducing a dynamic adjustment strategy, thereby improving the overall inspection efficiency and the practicality of path planning.
[0094] The embodiments that this application seeks to protect include:
[0095] A global path planning method for robot inspection of smart wind farm booster stations.
[0096] The steps include: Step S0: Pre-setting a set of empirical parameter values for improving the ant colony algorithm;
[0097] Step S1: Initialize the improved ant colony algorithm, setting the total number of ants m and the iteration threshold N;
[0098] Step S2: Perform the ant colony algorithm iterative process; at the beginning of each iteration, dynamically adjust the pheromone importance factor α, the heuristic function importance factor β, and the volatile factor ρ based on the historical evolution of the current global optimal solution;
[0099] Step S3 uses an improved ant colony algorithm to iterate, select the path with the lowest cost, update the pheromone concentration, calculate the contribution ratio of branches, and determine the termination condition of the iteration until the iteration number threshold is reached or other termination conditions are met.
[0100] Furthermore, the empirical parameter values for step S0 specifically include:
[0101] Pheromonic importance factor α: set to the empirical value α0;
[0102] The heuristic function importance factor β: set to the empirical value β0;
[0103] Volatility factor ρ: Set to the empirical value ρ0;
[0104] Empirical parameter values are selected based on the environmental characteristics of the wind farm's booster station and the requirements of the inspection tasks.
[0105] Furthermore, step S1 specifically involves:
[0106] Set the total number of ants to `m`;
[0107] Set a maximum iteration threshold `N` as one of the conditions for the algorithm to stop searching. If no better solution is found after exceeding this number of iterations, the algorithm is considered to be close to the optimal solution or trapped in a local optimum.
[0108] Construct a network topology diagram of the wind farm booster station, where each node represents the equipment or area to be inspected, and each edge represents the actual path connecting two nodes, with the corresponding distance or cost attached.
[0109] Initialize the pheromone matrix, ensuring a uniform initial pheromone concentration distribution across all paths or assigning certain prior information based on historical inspection data; initialize the global optimal solution record, including the current shortest path and its corresponding cost;
[0110] Each ant is randomly placed on a starting node and given an empty list of paths, ready to begin traversing the network to find the optimal path.
[0111] Further, step S2 includes: step S2.1: calculate the interval Δt between the current iteration number t and the most recent global optimal solution update iteration number t0; calculate the current iteration number `t`; obtain and record the iteration number `t0` when the global optimal solution was last found;
[0112] The time interval Δt = t - t0 is calculated based on the current iteration number `t` and the iteration number `t0` updated by the most recent global optimal solution; Δt reflects whether the current algorithm is trapped in a local optimum or the degree of effective search in the global scope.
[0113] Furthermore, step S2 includes: step S2.2: adjusting α, β, and ρ according to Δt and preset adjustment rules;
[0114] Step S2.2 in detail:
[0115] If Δ`t` is large and greater than the threshold, increase the value of the volatile factor ρ to accelerate the dissipation of pheromones, prompting ants to explore more unconsidered paths and break the existing local optimum trap; if Δ`t` is small and less than the threshold, maintain the volatile factor ρ or decrease the volatile factor ρ to maintain the current good search space convergence performance.
[0116] Adjust α and β based on Δ`t`, the quality change of the local solution, and the remaining number of iterations.
[0117] Furthermore, step S2 includes: Step S2.3: Updating the adjusted α, β, and ρ values for path selection by ants within the current iteration cycle; Step S2.3 specifically:
[0118] Set specific parameter adjustment rules, including linear increase and exponential decay, and calculate new α', β', and ρ' values based on the magnitude of Δ`t`;
[0119] The adjusted new parameter values α', β', and ρ' are applied to the ant path selection process in the current iteration cycle;
[0120] After each iteration cycle, check the algorithm performance to see if the new parameter settings have improved the search performance;
[0121] The updated parameters are used to continue the iterative process of the ant colony algorithm, including path selection, pheromone update, and branch contribution ratio calculation, until the preset iteration number threshold N is reached.
[0122] Furthermore, step S3 is detailed as follows:
[0123] Within the current iteration period t, each virtual ant selects a path according to the improved ant colony algorithm; each ant determines the probability of moving between nodes in the graph based on the adjusted parameter values α, β, and ρ; specifically, the probability of each node being selected as the next visit point consists of two parts: the selection probability based on pheromone intensity and the selection probability based on heuristic information.
[0124] All ants traverse the entire map network of the booster station, construct their own inspection paths, and calculate the corresponding path costs. After completing one round of iteration, the pheromone matrix is updated based on the path costs found by the ants. For the edges traversed by ants that have discovered better paths, the corresponding pheromone amount is increased. The pheromone on all edges is decayed according to the evaporation factor ρ.
[0125] Compare the costs of all paths found by all ants. If an ant finds a globally optimal path with a lower cost, update the globally optimal solution and record the iteration number t0 of this update.
[0126] Based on the importance and contribution of each branch in the optimal path, the weighting coefficient of its pheromone update is further calculated to guide the subsequent ants' search on key branches;
[0127] Check if the preset iteration count threshold N has been reached; if the threshold has not been reached, continue to the next iteration; if the threshold has been reached but the global optimal solution still does not meet the preset convergence criterion, extend the iteration count.
[0128] In each iteration, the three core parameters α, β, and ρ are continuously monitored and adjusted according to the dynamic adjustment strategy in step S2.2 to adapt to environmental changes and task requirements, enabling the algorithm to explore the global optimal path more effectively at different stages.
[0129] Furthermore, path costs include, but are not limited to, path length, expected inspection time, and costs of avoiding potentially hazardous areas.
[0130] The above description is merely one specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A global path planning method for robot inspection of smart wind farm booster stations, characterized in that, The steps include: Step S0: Pre-setting a set of empirical parameter values for improving the ant colony algorithm; Step S1: Initialize the improved ant colony algorithm, setting the total number of ants m and the iteration threshold N; Step S2: Perform the ant colony algorithm iterative process; at the beginning of each iteration, dynamically adjust the pheromone importance factor α, the heuristic function importance factor β, and the volatile factor ρ based on the historical evolution of the current global optimal solution; Step S3 uses an improved ant colony algorithm to iterate, select the path with the lowest cost, update the pheromone concentration, calculate the branch contribution ratio, and determine the iteration termination condition until the iteration number threshold is reached or other termination conditions are met. Step S2 includes: Step S2.1: Calculate the interval Δt between the current iteration number t and the most recent global optimal solution update iteration number t0; calculate the current iteration number `t`; obtain and record the iteration number `t` when the global optimal solution was found last time. o `; Update the iteration count `t` based on the current iteration count `t` and the most recent global optimal solution. o Calculate the time interval Δt=tt o Δ`t` reflects whether the current algorithm has gotten stuck in a local optimum or the degree to which it has searched effectively in the global scope; Step S2 includes: Step S2.2: Adjust α, β, and ρ according to Δt and the preset adjustment rules; Step S2.2 in detail: If Δ`t` is large and greater than the threshold, increase the value of the volatile factor ρ to accelerate the dissipation of pheromones, prompting ants to explore more unconsidered paths and break the existing local optimum trap; if Δ`t` is small and less than the threshold, maintain the volatile factor ρ or decrease the volatile factor ρ to maintain the current good search space convergence performance. α and β are adjusted based on Δ`t`, the quality change of the local solution, and the remaining number of iterations; Step S2 includes: Step S2.3: Update the adjusted α, β, and ρ values for ant path selection in the current iteration cycle; Step S2.3 in detail: Set specific parameter adjustment rules, including linear increase and exponential decay, and calculate new α', β', and ρ' values based on the magnitude of Δ`t`; The adjusted new parameter values α', β', and ρ' are applied to the ant path selection process in the current iteration cycle; After each iteration cycle, check the algorithm performance to see if the new parameter settings have improved the search performance; The updated parameters are used to continue the iterative process of the ant colony algorithm, including path selection, pheromone update, and branch contribution ratio calculation, until the preset iteration number threshold N is reached.
2. The global path planning method for robot inspection of smart wind farm booster stations according to claim 1, characterized in that, The empirical parameter values for step S0 specifically include: Pheromonic importance factor α: set to the empirical value α0; The heuristic function importance factor β: set to the empirical value β0; Volatility factor ρ: Set to the empirical value ρ0; Empirical parameter values are selected based on the environmental characteristics of the wind farm's booster station and the requirements of the inspection tasks.
3. The global path planning method for robot inspection of smart wind farm booster stations according to claim 2, characterized in that, Step S1 is as follows: Set the total number of ants to `m`; Set a maximum iteration threshold `N` as one of the conditions for the algorithm to stop searching. If no better solution is found after exceeding this number of iterations, the algorithm is considered to be close to the optimal solution or trapped in a local optimum. Construct a network topology diagram of the wind farm booster station, where each node represents the equipment or area to be inspected, and each edge represents the actual path connecting two nodes, with the corresponding distance or cost attached. Initialize the pheromone matrix, ensuring that the initial pheromone concentration is evenly distributed across all paths or that certain prior information is assigned based on historical inspection data. Initialize the global optimal solution record, including the current shortest path and its corresponding cost; Each ant is randomly placed on a starting node and given an empty list of paths, ready to begin traversing the network to find the optimal path.
4. The global path planning method for robot inspection of smart wind farm booster stations according to claim 1, characterized in that, Step S3 is detailed as follows: Within the current iteration period t, each virtual ant selects a path according to the improved ant colony algorithm; each ant determines the probability of moving between nodes in the graph based on the adjusted parameter values α, β, and ρ; specifically, the probability of each node being selected as the next visit point consists of two parts: the selection probability based on pheromone intensity and the selection probability based on heuristic information. All ants traverse the entire map network of the booster station, construct their own inspection paths, and calculate the corresponding path costs. After completing one round of iteration, the pheromone matrix is updated based on the path costs found by the ants. For the edges traversed by ants that have discovered better paths, the corresponding pheromone amount is increased. The pheromone on all edges is decayed according to the evaporation factor ρ. Compare the costs of all paths found by the ants. If an ant finds a globally optimal path with a lower cost, update the globally optimal solution and record the iteration number t. o ; Based on the importance and contribution of each branch in the optimal path, the weighting coefficient of its pheromone update is further calculated to guide the subsequent ants' search on key branches; Check if the preset iteration count threshold N has been reached; if the threshold has not been reached, continue to the next iteration; if the threshold has been reached but the global optimal solution still does not meet the preset convergence criterion, extend the iteration count. In each iteration, the three core parameters α, β, and ρ are continuously monitored and adjusted according to the dynamic adjustment strategy in step S2.2 to adapt to environmental changes and task requirements, enabling the algorithm to explore the global optimal path more effectively at different stages.
5. The global path planning method for robot inspection of smart wind farm booster stations according to claim 1, characterized in that, Path costs include, but are not limited to, path length, expected inspection time, and costs of avoiding potentially hazardous areas.
Citation Information
Patent Citations
Underwater robot time optimal path planning method based on interval optimization
CN110244759A
Unmanned bicycle path planning method based on ant colony algorithm and polar coordinate transformation
WO2018176595A1
