Aircraft route planning method and device combining rrt and aco algorithm

Abstract

The present disclosure relates to the technical field of aircraft route planning, and discloses an aircraft route planning method and device combining RRT and ACO algorithms, which comprises the following steps: performing space exploration by using the RRT algorithm to generate an initial route path; extracting a key node sequence and performing path optimization in a search space constituted by the key node sequence by using the ACO algorithm, and iteratively optimizing path smoothness and path length through a pheromone update mechanism to obtain an optimized route path; feeding back to the space exploration process of the RRT algorithm; and obtaining a final route path through multiple iterations. The present disclosure constructs a closed-loop interaction mechanism between RRT global exploration and ACO local optimization, so that the path optimization result obtained in the ACO stage can dynamically feedback and guide the random sampling and exploration direction of the RRT algorithm, realize self-sustained optimization of the algorithm in multiple iterations, and improve the efficiency and stability of obtaining a high-quality path in a complex airspace.

Description

Technical Field

[0001] This disclosure relates to the field of aircraft route planning technology, specifically to an aircraft route planning method and apparatus that combines RRT and ACO algorithms. Background Technology

[0002] In UAV path planning, a hybrid strategy is often used to balance planning speed and path quality. For example, a feasible path is first obtained through a fast search algorithm, and then an optimization algorithm is used to smooth and shorten the path.

[0003] However, such hybrid strategies typically employ a fixed, sequential process, lacking effective information exchange and coordination between the preceding and following stages. Specifically, the front-end search process cannot leverage the experience accumulated from back-end optimization, leading to aimless searching; while back-end optimization is constrained by the initial path framework provided by the front-end, making it difficult to fundamentally improve the path structure. This open-loop structure hinders the algorithm's ability to continuously improve itself through iteration, limiting its efficiency and stability in acquiring high-quality paths in complex spatial domains. Therefore, a closed-loop planning method capable of dynamic feedback and collaborative evolution between global exploration and local optimization is urgently needed. Summary of the Invention

[0004] In view of this, this disclosure provides an aircraft route planning method and apparatus that combines RRT and ACO algorithms to solve the problem of how to achieve a closed-loop planning method that dynamically feeds back and coordinates global exploration and local optimization, thereby improving the efficiency and stability of obtaining high-quality paths in complex airspace.

[0005] This disclosure provides an aircraft flight path planning method that combines RRT and ACO algorithms. The method includes: Based on the starting point, ending point, and airspace constraints of the aircraft's navigation mission, the RRT algorithm is used to explore space and generate an initial flight path. The key node sequence in the initial flight path is extracted, and the ACO algorithm is used to optimize the path in the search space formed by the key node sequence. The path smoothness and path length are iteratively optimized through the pheromone update mechanism to obtain the optimized flight path. The optimized flight path is fed back into the spatial exploration process of the RRT algorithm to guide the subsequent exploration direction of the RRT algorithm; The final flight path is obtained through multiple iterations via closed-loop interaction between the RRT algorithm and the ACO algorithm.

[0006] This disclosure also provides an aircraft flight path planning device that combines RRT and ACO algorithms, the device comprising: The path exploration module is used to explore space and generate an initial flight path based on the starting point, ending point and airspace constraints of the aircraft's flight mission using the RRT algorithm. The path optimization module is used to extract the key node sequence in the initial flight path and use the ACO algorithm to optimize the path in the search space formed by the key node sequence. The path smoothness and path length are iteratively optimized through the pheromone update mechanism to obtain the optimized flight path. The feedback guidance module is used to feed back the optimized flight path to the spatial exploration process of the RRT algorithm, so as to guide the subsequent exploration direction of the RRT algorithm; The iterative control module is used to perform multiple iterations through closed-loop interaction between the RRT algorithm and the ACO algorithm to obtain the final flight path.

[0007] This disclosure also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described aircraft route planning methods combining RRT and ACO algorithms.

[0008] This disclosure also provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the steps of any of the above-described aircraft route planning methods combining RRT and ACO algorithms.

[0009] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described aircraft route planning methods combining RRT and ACO algorithms.

[0010] In the aircraft route planning method and apparatus combining RRT and ACO algorithms disclosed in the above embodiments, a closed-loop interaction and feedback guidance mechanism between RRT global exploration and ACO local optimization is constructed. This achieves a closed-loop planning method for UAV path planning that dynamically feeds back and collaboratively evolves between global exploration and local optimization, thereby improving the efficiency and stability of obtaining high-quality paths in complex airspace. By feeding back the high-quality path information obtained from the ACO stage optimization to the sampling process in the RRT stage, the subsequent exploration direction can be adaptively guided towards historically high-quality areas, significantly improving the targeting and efficiency of the global search. Through multiple iterations, path length and smoothness can be simultaneously optimized while ensuring path feasibility and safety, thus providing a route planning method with high search efficiency, high path quality, and strong environmental adaptability for dense scenarios such as urban low-altitude areas. Attached Figure Description

[0011] To more clearly illustrate the embodiments of this disclosure, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating the aircraft flight planning method combining RRT and ACO algorithms provided in this embodiment of the disclosure; Figure 2 A schematic diagram of the structure of an aircraft route planning device combining RRT and ACO algorithms provided in an embodiment of this disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0013] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this disclosure.

[0014] In complex and dynamic scenarios such as low-altitude urban flight of drones, efficiently generating a flight path that simultaneously satisfies safety, economy, and flight smoothness has become a core technical challenge in the industry. The Rapidly-exploring Random Trees (RRT) algorithm and the Ant Colony Optimization (ACO) algorithm, due to their complementary advantages in rapid exploration in high-dimensional space and refined path optimization, have been considered for hybrid path planning.

[0015] However, current hybrid path planning methods in the industry often have the following problems: 1. The first hybrid path planning method adopts a fixed pipeline mode of "generate first, optimize later". The RRT and ACO run completely independently. The RRT conducts blind exploration without the guidance of historical experience, and the quality of its output initial path is random. The optimization process of ACO is strictly limited to the solution space defined by the initial path, making it difficult to break through the original structural defects, and the optimization effect has a clear ceiling.

[0016] 2. While the second hybrid path planning method attempts to expand the search range by introducing more flexible node sampling in the ACO optimization stage, its information flow remains unidirectional. That is, the ACO optimization process cannot feed back the path quality information obtained to the front-end RRT exploration module, causing each RRT plan to start from scratch with an unguided search. The algorithm as a whole lacks the ability to learn and iteratively evolve using historical optimization experience, facing challenges of performance instability and limited optimization efficiency in complex environments.

[0017] The common problem with the two methods mentioned above is that the global exploration and local optimization phases are disconnected, lacking effective coordination and feedback mechanisms. This prevents the algorithms from achieving a dynamic balance between efficiency and quality, and makes them ill-suited to the high demands of complex 3D scenarios such as urban low-altitude areas for path safety, economy, and smoothness. Therefore, there is an urgent need for an intelligent planning method that can achieve closed-loop interaction between exploration and optimization and possess self-evolution capabilities.

[0018] To address the aforementioned issues, various embodiments of this disclosure provide an aircraft route planning method combining RRT and ACO algorithms. The method includes: based on the starting point, ending point, and airspace constraints of the aircraft's flight mission, using the RRT algorithm for spatial exploration to generate an initial route path; extracting key node sequences from the initial route path and using the ACO algorithm to optimize the path within the search space formed by the key node sequences, iteratively optimizing path smoothness and path length through a pheromone update mechanism to obtain an optimized route path; feeding the optimized route path back to the RRT algorithm's spatial exploration process to guide the subsequent exploration direction of the RRT algorithm; and performing multiple iterations through a closed-loop interaction between the RRT and ACO algorithms to obtain the final route path.

[0019] It should be noted that, in the description of this disclosure, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this disclosure are used to distinguish similar objects and are not used to describe a particular order or sequence.

[0020] To enable those skilled in the art to better understand the present disclosure, the present disclosure will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating an aircraft flight planning method combining RRT and ACO algorithms provided in an embodiment of this disclosure. The method may include the following steps: Step S101: Based on the starting point, ending point and airspace constraints of the aircraft's navigation mission, the RRT algorithm is used to explore space and generate an initial navigation path.

[0022] In this embodiment, step S101 can be the initial stage of the entire planning process. By utilizing the characteristics of the RRT algorithm, a feasible path from a specified starting point to a specified ending point can be quickly discovered in a complex airspace environment.

[0023] The starting point and the ending point can be predetermined by the specific flight mission of the aircraft, and can respectively represent the starting position and the target position of the aircraft's flight.

[0024] Airspace constraints include at least one restrictive factor that can affect the flight safety and legality of an aircraft. For example, airspace constraints may include at least: areas of obstacles to be avoided, the airspace range of no-fly zones, or the maximum permitted flight altitude of an aircraft.

[0025] Furthermore, by constructing a random tree structure with the starting point as the root node and growing in the free space allowed by the airspace constraints, and by performing a series of random sampling, nearest neighbor search and collision detection operations through the RRT algorithm, the tree structure is gradually expanded until its branches are close enough to the endpoint, thereby connecting to form a preliminary flight route that meets basic safety requirements, i.e., the initial flight path.

[0026] Here, the path can ensure physical connectivity from the starting point to the end point, but there is usually room for further optimization in terms of path length and smoothness. For example, the path may have problems such as local non-smoothness and abrupt turns due to the randomness of the RRT algorithm.

[0027] Step S102: Extract the key node sequence from the initial navigation path, and use the ACO algorithm to optimize the path in the search space formed by the key node sequence. Iteratively optimize the path smoothness and path length through the pheromone update mechanism to obtain the optimized navigation path.

[0028] In this embodiment, this step can finely optimize the initial navigation path obtained in step S101.

[0029] Specifically, firstly, a sequence of key nodes that can characterize the overall direction and structural features of the initial flight path is extracted. Each key node in the sequence can include turning points where the path undergoes significant directional changes, as well as control points located near obstacle edges or no-fly zone boundaries, which are key locations constrained by airspace limitations.

[0030] By filtering the aforementioned key nodes, the search complexity of subsequent optimization algorithms can be significantly reduced while preserving the basic framework of the initial navigation path.

[0031] Subsequently, based on the aforementioned sequence of key nodes, a path optimization graph is constructed for the ACO algorithm to search. In this path optimization graph, key nodes are used as vertices, and feasible path segments connecting these key nodes are used as edges.

[0032] The ACO algorithm simulates the accumulation and volatilization mechanism of pheromones in ant colony foraging behavior, and performs multiple rounds of iterative search in the above path optimization graph to optimize the smoothness of the path and shorten the total length of the path.

[0033] In each iteration, the ant simulated by the ACO algorithm selects a direction of movement with a certain probability based on the current pheromone concentration on each edge of the path optimization graph and heuristic information reflecting prior knowledge such as the distance between nodes, thereby constructing a complete candidate path from the starting point to the end point.

[0034] Furthermore, after all individual ants have completed the construction of candidate paths, the ACO algorithm calculates the cost of each candidate path according to a pre-defined objective function, and performs pheromone updates based on the cost calculation results: 1. All pheromones on the edges are volatilized to a certain extent according to a preset volatilization rate to help the ACO algorithm escape local optima; 2. Increase the concentration of pheromones along the edges traversed by high-quality paths with relatively lower costs, so as to attract more ant individuals to choose these edges in subsequent iterations, forming a reinforcement effect of high-quality solutions.

[0035] Preferably, the pre-defined objective function can be the shortest flight path length.

[0036] Furthermore, through repeated iterations of construction, evaluation, and updating, pheromones accumulate continuously in high-quality path regions, and the algorithm gradually converges.

[0037] Finally, from all the candidate paths generated in the iterations, the path that satisfies the airspace constraints and has the best overall performance in terms of smoothness and length is selected as the output of this step, i.e., the optimized navigation path.

[0038] Here, optimizing the flight path significantly improves the economy and smoothness of flight while maintaining the feasibility of the initial path.

[0039] Step S103 involves feeding back the optimized navigation path to the space exploration process of the RRT algorithm to guide its subsequent exploration direction.

[0040] In this embodiment, this step can be used to transfer local optimization experience to the global search strategy.

[0041] Specifically, the optimized flight path output in step S102 is analyzed to extract information on key spatial regions traversed by the path in the airspace. These regions are identified and marked as guidance areas with high exploration value because they have generated historically high-quality paths.

[0042] Based on the aforementioned spatial region information, the generation strategy for random sampling points in the RRT algorithm is adjusted. In subsequent iterations or a new round of exploration, sampling is no longer completely uniform or random, but rather sampling points are generated within the aforementioned guiding region with a higher probability. This makes the expansion direction of the random tree structure of the RRT algorithm more targeted towards the guiding region, thereby transforming the global exploration behavior of the RRT algorithm from blind random exploration to experience-guided exploration. This allows the RRT algorithm to focus its exploration resources more intently on the spatial range that is more likely to include high-quality solutions.

[0043] Understandably, the feedback guidance mechanism described above in this step can improve the exploration efficiency and directionality of the RRT algorithm.

[0044] Step S104: The final navigation path is obtained by iterating multiple times through the closed-loop interaction between the RRT algorithm and the ACO algorithm.

[0045] In this embodiment, this step integrates and iteratively executes the above steps S101 to S103 to form a closed-loop optimization system of exploration, optimization and feedback. Each cycle can constitute an iterative process. The input of this iteration process is the guidance area information accumulated in the previous iteration, and the output is the navigation path updated after collaborative optimization in this round.

[0046] Specifically, in each iteration, based on the current spatial constraints and historical guidance information, step S101 is first executed, where the RRT algorithm performs an experience-guided space exploration to generate a new initial navigation path; then step S102 is executed, where the ACO algorithm refines and optimizes the key node sequence of the path to obtain a higher quality navigation path; finally, step S103 is executed, where the new path information obtained from this optimization is transformed into spatial guidance strategy feedback and updated to guide the exploration behavior of RRT in the next iteration.

[0047] Furthermore, the above iterative process will continue until a preset termination condition is met. The termination condition can be set as the maximum preset number of iterations or the performance of optimizing the navigation path (e.g., path length) in multiple consecutive iterations converging to a stable range.

[0048] When the iteration terminates, the best-performing flight path obtained in the last iteration or recorded in all iterations is output as the final flight path that satisfies all airspace constraints and achieves comprehensive optimality or near-optimality in terms of smoothness, length, and other indicators.

[0049] In the aircraft route planning method and apparatus combining RRT and ACO algorithms disclosed in the above embodiments, a closed-loop interaction and feedback guidance mechanism between RRT global exploration and ACO local optimization is constructed. This achieves a closed-loop planning method for UAV path planning that dynamically feeds back and collaboratively evolves between global exploration and local optimization, thereby improving the efficiency and stability of obtaining high-quality paths in complex airspace. By feeding back the high-quality path information obtained from the ACO stage optimization to the sampling process in the RRT stage, the subsequent exploration direction can be adaptively guided towards historically high-quality areas, significantly improving the targeting and efficiency of the global search. Through multiple iterations, path length and smoothness can be simultaneously optimized while ensuring path feasibility and safety, thus providing a route planning method with high search efficiency, high path quality, and strong environmental adaptability for dense scenarios such as urban low-altitude areas.

[0050] In one possible implementation of the above embodiment, before step S101, the method further includes: performing three-dimensional mesh modeling of the flight airspace, including: discretizing the flight airspace along the three-dimensional coordinate axes into a grid with a side length of... Uniform cubic blocks; construct a three-dimensional array to describe the passable state of each cubic block. And spatial constraints are formed based on the three-dimensional array.

[0051] In this embodiment, the purpose of three-dimensional mesh modeling is to transform a continuous, physical flight airspace into a discrete, digital model composed of regular units, so as to perform state management and path search.

[0052] Specifically, first, the basic spatial resolution is set, which is the fixed side length of the cube block in the X, Y, and Z coordinate axes, and can be denoted as... .

[0053] Next, the entire three-dimensional flight airspace to be studied is uniformly divided along the three coordinate axes using the aforementioned fixed side lengths, thus obtaining a series of side lengths. Uniform cubic blocks.

[0054] For example, suppose the size of the city's low-altitude airspace is The side length of each block can be set to... Then you can get Each block.

[0055] Each block can be represented by an index value. The unique identifier, and the representation of the entire spatial domain using a three-dimensional array, can be shown in the following formula:

[0056] in, , It is a set of non-negative integers.

[0057] Here, the set of all blocks constitutes a discretized coverage of the entire flight airspace, which can be represented by a three-dimensional array. The array provides a unified representation and management system. Each element in the array corresponds to a specific block, which is used to store and describe the real-time passability status of that block.

[0058] For example, real-time passability status can include, but is not limited to: idle (free to pass), occupied (occupied by obstacles or aircraft and cannot pass), or planned to be occupied (soon to be occupied and needs to be avoided during planning).

[0059] Here, the state model based on the three-dimensional array described above is a specific digital representation of the spatial constraints relied upon in the subsequent path planning process. For example, during path planning, it is required that the state of all blocks traversed by the navigation path must be idle.

[0060] Furthermore, in order to improve the path planning accuracy in complex areas while ensuring computational efficiency, dynamic mesh refinement technology can also be used.

[0061] Preferably, in areas where the aircraft is expected to be in dense airspace or where there are complex static obstacles, the local area can be further subdivided automatically or under the trigger of preset rules.

[0062] For example, the original side length Shrink to ,in, .

[0063] Here, by using the aforementioned method of local subdivision, a more refined environmental model can be established in key areas, thereby supporting more accurate complex flight paths and avoiding the huge computational overhead caused by high-resolution modeling of the entire airspace.

[0064] In the aircraft route planning method and apparatus combining RRT and ACO algorithms in the above embodiments of this disclosure, by discretizing the continuous flight airspace into a regular three-dimensional mesh model, a precise, structured, and digital expression of complex airspace constraints such as obstacles and no-fly zones is achieved. This provides a unified, clear, and computable environmental state input for subsequent path planning algorithms, fundamentally ensuring the physical feasibility and safety of the planned path. By introducing a dynamic mesh refinement mechanism, local mesh densification can be automatically performed in densely populated flight areas or around complex terrain, thereby achieving fine-grained control of the path orientation in key areas without significantly increasing the global computational load, effectively improving the planning accuracy of the route at complex constraint boundaries.

[0065] In one possible implementation of step S101 above, the RRT algorithm is used for space exploration to generate an initial navigation path, including: Initialize a search tree with the starting point as the root node; A sampling point is randomly generated within the passable space defined by the spatial constraints. Determine the node closest to the sampling point among all nodes in the search tree; Expand a new node by a preset step size from the nearest node toward the sampling point; Check if the path segment from the nearest node to the new node satisfies the spatial constraints; if it does, add the new node to the search tree as a child node of the nearest node. By repeatedly executing an iterative process that includes randomly generating sampling points, determining the nearest node, expanding the preset step size, checking, and adding, until the distance between the new node and the endpoint is less than a preset threshold, the path from the root node to the new node at this point is taken as the initial navigation path.

[0066] In this embodiment, the RRT algorithm is initialized to create a starting point specified by the aircraft's flight mission. Search tree with root node , where the search tree It will continue to grow in subsequent iterations, and its nodes can represent feasible locations in the spatial domain, and its edges can represent safe path segments.

[0067] Furthermore, the RRT algorithm enters an iterative exploration process, performing the following steps in each iteration: Step a1, under the spatial constraints (i.e., the three-dimensional array) Within the passable space defined by the set of blocks marked as free in the (), a sampling point is randomly generated. .

[0068] Step a2, traverse the current search tree For all nodes, quickly find the distance using a distance metric function (such as Euclidean distance). The most recent existing tree node .

[0069] Step a3, from node Start, following the directions Expand in the direction by a preset step size Generate a new node .

[0070] Here, the expansion process can be represented by the following formula:

[0071] in, Representing vectors The modulus (i.e., length) of the new node ensures that the new node... Located from point to In the direction, and with The distance is the preset step size. .

[0072] Step a4, generate from arrive The path segment is identified, and it is checked whether the path segment meets all spatial constraints. If it meets all spatial constraints, the path segment is determined to be feasible; otherwise, it is determined to be infeasible.

[0073] Here, spatial constraints include at least the following three categories: The path cannot traverse grids that are occupied or planned to be occupied; the Z coordinate (i.e., altitude) of each node on the path must meet the flight altitude limit; and the path must be a continuous and unbranched single path in the topology.

[0074] Specifically, the Z-coordinate of each node on the path satisfies the flight altitude limit, as shown in the following formula:

[0075] in, Indicates the first in the path Flight altitude at each node Indicates the minimum flight altitude. Indicates the maximum flight altitude.

[0076] Furthermore, the path must be a topologically continuous and unbranched single path, signifying that the aircraft's flight path must be continuous across different dimensions. This means that each grid point can only have one predecessor and one successor. The path continuity constraint can be defined as follows:

[0077] in, Indicates the first in the path Each grid node corresponds to a specific location in the path coordinate sequence; It is a binary variable used to represent the slave node. To the node Does a valid connection exist? If so, then ,otherwise This constraint requires each node There must be one and only one predecessor node in the path. This ensures the uniqueness and continuity of the path in the three-dimensional grid space.

[0078] Step a5, if arrive If the path segment verification passes, then... Add to search tree and will Set as The child nodes, at this point in the search tree The branch implements expansion.

[0079] The above steps will be repeated, allowing the search tree to gradually spread from the starting point outwards into the surrounding passable space.

[0080] The termination condition of the RRT algorithm is determined by the endpoint. Trigger: When generated in a certain iteration and When the spatial distance between the start and end points is less than a preset threshold, the RRT algorithm determines that a channel connecting the start and end points has been found.

[0081] At this time, by means of Backtracking to the root node This allows us to extract the unique branch connecting the start and end points in the search tree, and the sequence of nodes traversed by this branch is output as the initial navigation path.

[0082] Furthermore, the complexity of the RRT algorithm depends on the number of samples and the efficiency of the nearest neighbor search. If efficient data structures such as KD-trees are used, the overall complexity can be approximated as... .

[0083] In the aircraft route planning method and apparatus combining RRT and ACO algorithms in the above embodiments of this disclosure, by executing the iterative exploration process of the RRT algorithm within a three-dimensional meshed airspace model, the global coverage and directional expansion mechanism of random sampling can be effectively utilized to quickly generate an initial feasible path from the starting point to the ending point in an environment with complex obstacles and altitude constraints. This significantly improves the efficiency of obtaining reliable input paths necessary for high-quality optimization algorithms. During each expansion, the path segment is verified in real time to ensure that it meets the constraints of obstacle avoidance, continuity, and flight altitude, ensuring that the generated initial path meets airspace safety requirements and improving the reliability of path planning.

[0084] In one possible implementation of step S102 above, the key node sequence in the initial navigation path is extracted, including: Extract the sequence of key nodes from the initial flight path, where the sequence of key nodes includes key nodes where the path direction changes, as well as key nodes located at the edge of obstacles and / or the boundary of no-fly zones as defined by airspace constraints. Based on the sequence of key nodes, a path optimization graph for ACO algorithm search is constructed, where the vertices of the path optimization graph are key nodes and the edges are candidate path segments connecting the key nodes.

[0085] In this embodiment, the node sequence on the initial navigation path is analyzed to identify two types of key nodes that play a decisive role in the path morphology and feasibility: path turning nodes and constraint neighbor nodes.

[0086] Among them, path turning nodes can be nodes where the path changes direction in three-dimensional space, and these nodes can constitute the basic direction of the path; constraint neighbor nodes can be nodes located near the edge of the obstacle surface or the boundary of the no-fly zone as defined by the airspace constraint conditions. These nodes are used to identify key spatial locations that the path must approach or detour to meet safety constraints such as obstacle avoidance.

[0087] By extracting the two types of key nodes mentioned above, a key node sequence can be obtained, which greatly simplifies the complexity of the original path data while fully preserving its topological structure, key turning points, and key positions that interact with complex spatial constraints.

[0088] Furthermore, based on the aforementioned key nodes, a path optimization graph is constructed (e.g. In this path optimization graph, each critical node is represented by a vertex (…). ); establish edges between sequentially adjacent key nodes, and between other key nodes that can be connected according to preset rules (e.g., within a certain Euclidean distance and line-of-sight range). ).

[0089] Each edge represents a candidate path segment connecting two key nodes. Geometrically, these candidate path segments can be straight lines or curves that meet the minimum turning radius, and they need to undergo fast collision detection during construction to ensure that they themselves meet basic spatial constraints, thereby guaranteeing that any path searched on the graph by the ACO algorithm is physically flyable.

[0090] In the aircraft route planning method and apparatus combining RRT and ACO algorithms in the above embodiments of this disclosure, the search space of the optimization algorithm is significantly compressed by extracting key nodes representing the path direction and constraint boundaries from the initial flight path. This significantly reduces the computational complexity of the subsequent ACO algorithm while preserving the core structural information of the path. By incorporating safety-critical locations such as obstacle edges and no-fly zone boundaries into the key nodes, the optimization process is guided within a solution space that ensures safety, thereby improving the reliability and feasibility of the optimized path.

[0091] In one possible implementation of step S102 above, the ACO algorithm is used to optimize the path in the search space formed by the sequence of key nodes. The path smoothness and path length are iteratively optimized through a pheromone update mechanism to obtain the optimized navigation path, including: Initialize the pheromone concentration of each edge in the path optimization graph; Perform multiple iterations, and output the optimized flight path as the path that satisfies the airspace constraints and has the lowest cost after multiple iterations. Each iteration includes the following steps: For each ant in the ant colony, based on the probability determined by the current pheromone concentration of each edge on the path optimization graph and heuristic information, subsequent nodes are selected step by step from the starting point to the end point in the path optimization graph, thereby constructing a corresponding candidate optimization path for each ant. Based on the preset objective function, calculate the cost of each candidate optimization path. The objective function should include at least a measure of the path length. The pheromone concentration on each edge in the path optimization graph is updated so that the existing pheromone concentration on each edge evaporates at a preset evaporation rate. The edges traversed are then enhanced with pheromone based on the cost of each candidate optimization path, with edges corresponding to paths with lower costs receiving higher enhancements.

[0092] In this embodiment, during the initialization process, each edge in the path optimization graph (which can be used) is optimized. This indicates that the vertices are connected. and Initialize its pheromone concentration Set the maximum number of iterations to The initial pheromone concentration of all edges can be set to the same positive constant. .

[0093] In addition, initialize parameters in the ACO algorithm, such as the number of ants. Maximum number of iterations pheromone initial value pheromone volatilization rate And parameters that control the heuristic and pheromone weights. and There are also constants. Used to calculate pheromone increments.

[0094] Furthermore, the algorithm performs multiple iterations, in each iteration... This process can include the following three core steps: Step b1, ant path construction.

[0095] Here, release Only one ant performs a parallel search. Each ant... ( Starting from the vertex representing the starting point, and based on a state transition probability formula, the next vertex is selected step by step on the path optimization graph until the vertex representing the ending point is reached, thus constructing a complete candidate optimization path. .

[0096] When the ant chooses the next vertex At that time, based on the current vertex Domain vertex set Current pheromone concentration on each edge and heuristic information Perform the calculation.

[0097] Select Vertex probability It can be represented by the following formula:

[0098] in, Representing an edge The concentration of pheromones on the surface For heuristic information, and Adjust the importance of pheromones and heuristic information respectively; Represents a node The neighborhood set, Indicates the first The complete path constructed by an ant in this iteration is an ordered set of nodes traversed from the starting point to the ending point.

[0099] Preferably, heuristic information Can be located as a vertex and The probability formula above is the reciprocal of the Euclidean distance between the two edges. Therefore, it can be ensured that ants tend to choose edges with high pheromone concentration and short distances, while maintaining a certain degree of random exploration.

[0100] Step b2, path cost evaluation.

[0101] Here, after all ants have completed path construction, the path for each ant is calculated according to a preset objective function. Total cost .

[0102] Preferably, in this embodiment, the main optimization objective is the path length. Therefore, the objective function can be defined as the sum of the lengths of all edges on the path, as shown in the following formula:

[0103] Among them, the cost The smaller the value, the better it represents a path. The better.

[0104] Step b3, pheromone update.

[0105] Here, pheromone renewal can include two processes: evaporation and enhancement. Evaporation can refer to simulating the natural evaporation of pheromones over time to prevent their unlimited accumulation. For example, pheromones on all edges can be evaporated at a preset rate. Attenuation, volatility The range of values ​​is Here, the larger It can accelerate the evaporation rate to improve search diversity, and the smaller the evaporation rate... This can enhance convergence stability.

[0106] Preferably, pheromone evaporation can be represented by the following formula:

[0107] Furthermore, enhancement can characterize the pheromone reward applied to the edges traversed by a path based on the quality of the paths found by all ants in the current iteration.

[0108] Preferably, pheromone enhancement can be achieved using the following formula:

[0109] in, It is a pheromone constant factor that can be used to control the magnitude of the pheromone increment per unit path cost. This represents the total number of ants involved in constructing the path in each iteration, i.e., the total number of ants generated in this round. The system identifies several paths and updates the pheromone concentration on all paths accordingly.

[0110] Furthermore, by adding the increments of all ants to the volatile pheromones, we obtain the initial pheromone concentration for the next iteration, as shown in the following formula:

[0111] By repeating steps b1 to b3 above, the pheromone concentration on high-quality paths continuously increases as more ants choose them and receive greater enhancement, which in turn attracts more subsequent ants, forming a positive feedback loop that guides the entire ant colony's search to gradually converge to a high-quality solution region.

[0112] When the number of iterations reaches the maximum The ACO algorithm terminates when the cost of the optimal path no longer improves after multiple iterations. At this point, from all candidate paths generated in the iterations, the path that satisfies the spatial constraints and has the highest total cost is selected. The path with the shortest minimum length is output as the final optimized navigation path. .

[0113] In the aircraft route planning method and apparatus combining RRT and ACO algorithms in the above embodiments of this disclosure, by combining pheromone evaporation and a path cost-based reverse enhancement update mechanism, the algorithm can both focus on searching historical high-quality path areas and maintain the ability to explore unexplored areas, thereby avoiding the optimization process from getting trapped in local optima. By using path length as the core cost for pheromone reward, the ant colony search is driven to continuously approach shorter and smoother flight paths, thereby improving the economy and flyability of the flight path while ensuring compliance with airspace constraints.

[0114] In one possible implementation of step S103 above, the optimized navigation path is fed back to the space exploration process of the RRT algorithm to guide the subsequent exploration direction of the RRT algorithm, including: Based on the optimized flight path, determine one or more spatial blocks in the airspace that the optimized flight path passes through; Based on the optimization results of the navigation path, the random sampling strategy of the RRT algorithm is adjusted to increase the probability of generating random sampling points in one or more space blocks in the next round of space exploration.

[0115] In this embodiment, the optimized navigation path output in step S102 is... Perform the analysis.

[0116] Specifically, for Calculate the grid index to which each point belongs. The set of blocks corresponding to these indices represents the physical space region occupied or adjacent to the optimized navigation path. These regions are marked as high-value guidance regions or historically high-quality path regions. These regions are considered more likely to contain other parts of a globally better solution because they generated the optimal path in the current iteration cycle.

[0117] In the next round or iteration of space exploration in step S101, the RRT algorithm adaptively adjusts its random sampling point generation strategy based on the above analysis results. The adjustment can aim to make the newly generated random sampling points... There is a higher probability that it will appear in the aforementioned high-value guidance area.

[0118] Preferably, a typical implementation can be non-uniform probability sampling, including: assigning a basic sampling weight to each grid block in the entire spatial domain, while significantly increasing the weight of blocks marked as high-value guiding regions. In the generation... First, a grid block is randomly selected based on the weight distribution, and then a three-dimensional coordinate point is uniformly and randomly generated within that block.

[0119] In this way, the random exploration process of the RRT algorithm is transformed from a completely uniform blind exploration to a biased exploration guided by historical optimization experience.

[0120] Through the aforementioned feedback mechanism, the RRT algorithm can focus more on potentially valuable spatial regions during subsequent explorations, thus more efficiently discovering initial paths with superior structures. This improved initial path is then input into the ACO module for further optimization, potentially generating even higher-quality optimized paths. These new optimized paths are then fed back to further refine the RRT algorithm's random sampling point generation strategy. This cycle repeats, forming a positive feedback loop between the RRT and ACO algorithms.

[0121] Further analysis of the algorithm's complexity reveals that the main computational load in the RRT algorithm stage is concentrated on random sampling and nearest neighbor search. If an efficient data structure, such as a KD-tree, is used, the average complexity can be reduced to [missing value]. The complexity of the ACO phase algorithm is related to the number of ants. Path length and number of iterations Proportional, the overall complexity can be approximated as Therefore, the overall complexity of the RRT-ACO algorithm is O(n log n). By adjusting the parameters, a balance can be achieved between computational efficiency and path quality.

[0122] In the aircraft route planning method and apparatus combining RRT and ACO algorithms described in the above embodiments of this disclosure, by adjusting the random sampling probability distribution of the RRT algorithm based on the guidance signal, the algorithm's sampling points can adaptively focus on areas with better historical performance, significantly improving the targeting and efficiency of global exploration and avoiding ineffective searches in low-potential areas. By establishing a feedback mechanism from local optimization to global exploration, the hybrid algorithm of RRT and ACO possesses the ability to autonomously learn and continuously improve from the iteration history, thereby achieving a synergistic improvement in the overall performance of path planning and further enhancing the overall quality and efficiency of path planning.

[0123] In one embodiment, an aircraft route planning device 200 combining RRT and ACO algorithms is provided. This aircraft route planning device 200 corresponding one-to-one with the aircraft route planning method combining RRT and ACO algorithms in the above embodiments is provided. Figure 2 As shown, the device includes: The path exploration module 201 is used to explore space based on the starting point, ending point and airspace constraints of the aircraft's navigation mission, and generate an initial flight path using the RRT algorithm. The path optimization module 202 is used to extract the key node sequence in the initial route path and use the ACO algorithm to optimize the path in the search space formed by the key node sequence. The path smoothness and path length are iteratively optimized through the pheromone update mechanism to obtain the optimized route path. The feedback guidance module 203 is used to feed back the optimized flight path to the spatial exploration process of the RRT algorithm in order to guide the subsequent exploration direction of the RRT algorithm; The iterative control module 204 is used to perform multiple iterations through the closed-loop interaction between the RRT algorithm and the ACO algorithm to obtain the final route path.

[0124] In one embodiment, the apparatus further includes a three-dimensional modeling module 205, wherein: 3D modeling module 205 is used to perform 3D mesh modeling of the flight airspace, including: discretizing the flight airspace along the 3D coordinate axes into a grid with side lengths of... Uniform cubic blocks; construct a three-dimensional array to describe the passable state of each cubic block. And spatial constraints are formed based on the three-dimensional array.

[0125] In one embodiment, the path exploration module 201 is specifically used to initialize a search tree with the starting point as the root node; A sampling point is randomly generated within the passable space defined by the spatial constraints. Determine the node closest to the sampling point among all nodes in the search tree; Expand a new node by a preset step size from the nearest node toward the sampling point; Check if the path segment from the nearest node to the new node satisfies the spatial constraints; if it does, add the new node to the search tree as a child node of the nearest node. By repeatedly executing an iterative process that includes randomly generating sampling points, determining the nearest node, expanding the preset step size, checking, and adding, until the distance between the new node and the endpoint is less than a preset threshold, the path from the root node to the new node at this point is taken as the initial flight path.

[0126] In one embodiment, the path optimization module 202 is specifically used to extract a sequence of key nodes from the initial flight path, wherein the sequence of key nodes includes key nodes where the path direction changes, and key nodes located at the edge of obstacles and / or the boundary of no-fly zones defined by airspace constraints. Based on the sequence of key nodes, a path optimization graph for ACO algorithm search is constructed, where the vertices of the path optimization graph are key nodes and the edges are candidate path segments connecting the key nodes.

[0127] In one embodiment, the path optimization module 202 is specifically used to initialize the pheromone concentration of each edge in the path optimization graph; Perform multiple iterations, and output the optimized flight path as the path that satisfies the airspace constraints and has the lowest cost after multiple iterations. Each iteration includes the following steps: For each ant in the ant colony, based on the probability determined by the current pheromone concentration of each edge on the path optimization graph and heuristic information, subsequent nodes are selected step by step from the starting point to the end point in the path optimization graph, thereby constructing a corresponding candidate optimization path for each ant. Based on the preset objective function, calculate the cost of each candidate optimization path. The objective function should include at least a measure of the path length. The pheromone concentration on each edge in the path optimization graph is updated so that the existing pheromone concentration on each edge evaporates at a preset evaporation rate. The edges traversed are then enhanced with pheromone based on the cost of each candidate optimization path, with edges corresponding to paths with lower costs receiving higher enhancements.

[0128] In one embodiment, the feedback guidance module 203 is specifically used to determine one or more spatial blocks in the airspace that the optimized flight path passes through, based on the optimized flight path. Based on the optimization results of the flight path, the random sampling strategy of the RRT algorithm is adjusted to increase the probability of generating random sampling points in one or more spatial blocks in the next round of space exploration.

[0129] It should be noted that the aircraft route planning device combining RRT and ACO algorithms provided in the above embodiments is only illustrated by the division of the above-described program modules when implementing the corresponding aircraft route planning method combining RRT and ACO algorithms. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the above device can be divided into different program modules to complete all or part of the processing described above. In addition, the device provided in the above embodiments and the corresponding Figure 1 The embodiments of the methods shown belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0130] This disclosure also provides an electronic device having the above-described features. Figure 2 The shown is an aircraft route planning device that combines RRT and ACO algorithms.

[0131] Figure 3This is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure.

[0132] The following is a detailed reference. Figure 3 The diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present disclosure. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 301, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 302 or a program loaded from memory 308 into random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device. The processor 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0133] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0134] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from memory 308, or installed from ROM 302. When the computer program is executed by processor 301, it performs the functions defined in the aircraft route planning method combining RRT and ACO algorithms of embodiments of this disclosure.

[0135] Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0136] This disclosure also provides a computer-readable storage medium in which the methods described in this disclosure can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code downloaded over a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the aircraft route planning method combining RRT and ACO algorithms shown in the above embodiments is implemented.

[0137] A portion of this disclosure can be applied to computer program products, such as computer program instructions, which, when executed by a computer, can invoke or provide methods and / or technical solutions according to this disclosure through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, and installation package files. Accordingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions; the computer compiling the instructions and then executing the corresponding compiled program; the computer reading and executing the instructions; or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0138] Although embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. An aircraft flight path planning method combining RRT and ACO algorithms, characterized in that, The method includes: Based on the starting point, ending point, and airspace constraints of the aircraft's navigation mission, the RRT algorithm is used to explore space and generate an initial flight path. The key node sequence in the initial flight path is extracted, and the ACO algorithm is used to optimize the path in the search space formed by the key node sequence. The path smoothness and path length are iteratively optimized through the pheromone update mechanism to obtain the optimized flight path. The optimized flight path is fed back into the spatial exploration process of the RRT algorithm to guide the subsequent exploration direction of the RRT algorithm; The final flight path is obtained through multiple iterations via closed-loop interaction between the RRT algorithm and the ACO algorithm.

2. The method according to claim 1, characterized in that, Before generating the initial flight path, the method further includes: Perform three-dimensional mesh modeling of the flight airspace, including: discretizing the flight airspace along the three-dimensional coordinate axes into meshes with side lengths of... Uniform cubic blocks; construct a three-dimensional array to describe the passable state of each cubic block. The spatial constraints are formed based on the three-dimensional array.

3. The method according to claim 1 or 2, characterized in that, The process of using the RRT algorithm for space exploration to generate an initial flight path includes: Initialize a search tree with the starting point as the root node; A sampling point is randomly generated within the passable space defined by the spatial constraints. In each node of the search tree, determine the node closest to the sampling point; Extend a preset step size from the nearest node toward the sampling point to generate a new node; Check whether the path segment from the nearest node to the new node satisfies the spatial constraint condition; if it does, add the new node to the search tree as a child node of the nearest node; By repeatedly executing the iterative process that includes randomly generating sampling points, determining the nearest node, expanding the preset step size, checking, and adding, until the distance between the new node and the endpoint is less than a preset threshold, the path from the root node to the new node at this time is taken as the initial flight path.

4. The method according to claim 3, characterized in that, The extraction of the key node sequence from the initial flight path includes: Extract a sequence of key nodes from the initial flight path, wherein the sequence of key nodes includes key nodes where the path direction changes, and key nodes located at the edge of obstacles and / or the boundary of no-fly zones as defined by the airspace constraints. Based on the sequence of key nodes, a path optimization graph for the ACO algorithm search is constructed, wherein the vertices of the path optimization graph are the key nodes, and the edges are candidate path segments connecting the key nodes.

5. The method according to claim 4, characterized in that, The ACO algorithm is used to optimize the path in the search space formed by the sequence of key nodes. The path smoothness and path length are iteratively optimized through a pheromone update mechanism to obtain the optimized flight path, including: Initialize the pheromone concentration of each edge in the path optimization graph; Perform multiple iterations, and output the optimized flight path as the path that satisfies the airspace constraints and has the lowest cost after multiple iterations. Each iteration includes the following steps: For each ant in the ant colony, based on the probability determined by the current pheromone concentration and heuristic information of each edge on the path optimization graph, subsequent nodes are selected step by step from the starting point to the end point in the path optimization graph, thereby constructing a corresponding candidate optimization path for each ant. The cost of each candidate optimization path is calculated based on a preset objective function, wherein the objective function includes at least a measure of the path length. The pheromone concentration on each edge in the path optimization graph is updated so that the existing pheromone concentration on each edge evaporates at a preset evaporation rate. The edges traversed are then enhanced with pheromone based on the cost of each candidate optimization path, with the edges corresponding to paths with lower costs receiving higher enhancement.

6. The method according to claim 1, characterized in that, The process of feeding back the optimized flight path to the RRT algorithm for spatial exploration, in order to guide the subsequent exploration direction of the RRT algorithm, includes: Based on the optimized flight path, one or more spatial blocks in the airspace are determined by the optimized flight path. Based on the optimization results of the optimized flight path, the random sampling strategy of the RRT algorithm is adjusted so that the probability of generating random sampling points in one or more space blocks is increased in the next round of space exploration.

7. An aircraft flight path planning device combining RRT and ACO algorithms, characterized in that, The device includes: The path exploration module is used to explore space and generate an initial flight path based on the starting point, ending point and airspace constraints of the aircraft's flight mission using the RRT algorithm. The path optimization module is used to extract the key node sequence in the initial flight path, and use the ACO algorithm to optimize the path in the search space formed by the key node sequence. The path smoothness and path length are iteratively optimized through the pheromone update mechanism to obtain the optimized flight path. The feedback guidance module is used to feed back the optimized flight path to the spatial exploration process of the RRT algorithm, so as to guide the subsequent exploration direction of the RRT algorithm; The iterative control module is used to perform multiple iterations through the closed-loop interaction between the RRT algorithm and the ACO algorithm to obtain the final flight path.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the aircraft route planning method combining RRT and ACO algorithms as described in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the aircraft route planning method combining the RRT and ACO algorithms as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the aircraft route planning method combining RRT and ACO algorithms as described in any one of claims 1 to 6.

Images