Route planning method based on multi-dimensional information characterization and evolution calculation

Abstract

The invention provides an air route planning method based on multi-dimensional information representation and evolution calculation, and belongs to the technical field of unmanned aerial vehicle air route planning. Firstly, task demand and scene analysis is carried out, multi-dimensional information is processed, threat prediction is carried out, elevation information is complemented through neighborhood interpolation based on topographic data, and a three-dimensional image is obtained; smoothing the terrain to construct a minimum safe flight curved surface, determining a detection blind area based on a terrain shielding effect, delimiting an initial task feasible area, and dynamically predicting a threat target position to generate a dynamic threat situation map; the method comprises the following steps: constructing a multi-target air route optimization model, performing model solution based on an improved ant colony algorithm, designing and differentially maintaining composite environment pheromones, generating an initial feasible air route through a feasible air route increment construction method considering space-time constraints, and dynamically updating the pheromones to complete final air route planning; according to the method, the problems of poor air route collaboration, low concealment, insufficient feasibility and the like caused by space-time and performance constraints in a threat environment can be solved.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) route planning technology, specifically, it relates to a route planning method based on multidimensional information representation and evolutionary calculation. Background Technology

[0002] When drone swarms fly at low altitudes towards their designated mission areas, they face the combined effects of uncertain countermeasures and a complex and ever-changing external environment. Therefore, proactive planning measures are necessary to ensure the robustness and safety of flight paths. The interactions between different environmental elements and flight path planning are summarized below: Terrain obstacles: Natural barriers formed by complex terrain (such as mountains, canyons and forests) provide possibilities for low-altitude penetration missions. Their shielding effect on threats needs to be considered in the planning by introducing three-dimensional terrain data and corresponding optimization models, so as to ensure the rationality and safety of route planning.

[0003] Countering Threats: The deployment of different types of enemy defense systems is dynamic, and the threat intensity changes over time, thus placing higher demands on the stealth of flight routes. To address this uncertainty, it is necessary to predict the distribution of potential threats over a future period and incorporate the results into the route optimization process to ensure mission safety.

[0004] Environmental factors: Random weather factors in the environment (such as wind speed and wind direction) also have an important impact on flight routes. These can be treated as no-fly zones by modeling local meteorological conditions, thereby reducing flight threats and improving mission reliability.

[0005] Taking into account the effects of the above interactions, the planned flight path typically needs to meet the following requirements: High stealth: The key objective of route planning is to ensure low detectability and reduce energy consumption of flight routes while meeting the maneuverability requirements of different types of platforms. To this end, it is necessary to make full use of low-altitude flight, complex terrain concealment, and radar blind spots to reduce the probability of detection of the route segment, while actively avoiding high-threat areas such as destructive weapons and no-fly zones.

[0006] High safety: Low-altitude flight requires dealing with complex environmental threats, including terrain obstacles and weather conditions. Route planning must select areas with ample available maneuvering airspace to ensure flight safety and provide room for response to emergencies.

[0007] High degree of coordination: To ensure stealth, drone swarms typically employ evasive maneuvers, approaching targets from different directions. During flight, adequate safety margins must be maintained in both time and space to ensure coordination within the swarm and across different flight paths.

[0008] In summary, it can be seen that during the low-altitude flight of drone swarms, they inevitably face complex mission environments, including diverse threat deployments, varied terrain, and strict mission constraints. These factors significantly increase the difficulty of route planning. Summary of the Invention

[0009] To address the aforementioned issues, this invention proposes a route planning method based on multidimensional information representation and evolutionary computation. By integrating multidimensional information such as terrain, threats, and mission requirements, it first generates a safe flight surface using terrain processing and constructs an initial mission space based on threat blind spots. Then, by combining terrain slope changes to predict the potential movement trends of air defense measures, a multi-objective route optimization model under multiple constraints is constructed. To obtain feasible routes that meet spatiotemporal coordination requirements, a solution method based on incremental route construction is proposed. Through differentiated design of the environment and heuristic information structure, as well as the adoption of co-evolutionary strategies such as improved ant colony algorithms, the search efficiency and route planning quality are significantly improved, while effectively avoiding high-threat areas. This solves the problems of poor route coordination, low concealment, and insufficient feasibility caused by spatiotemporal and performance constraints under threat environments.

[0010] This invention is achieved through the following technical solution: A route planning method based on multidimensional information representation and evolutionary computation: The method specifically includes the following steps: Step 1: Task requirements and scenario analysis, clarify the interaction between terrain obstacles, counter-threat, environmental factors and route planning, define the stealth, security and coordination requirements that the route must meet, and abstract the unmanned swarm, target, threat and environmental elements into an initial state set, a target attribute set, a waypoint information set, a threat set, an environmental situation set and a constraint set. Step 2: Multidimensional information processing and threat prediction. Based on terrain data and by supplementing elevation information through neighborhood interpolation, the terrain is smoothed to construct the minimum safe flight surface. Based on the terrain occlusion effect, the detection blind zone is determined and the initial mission feasible area is delineated. At the same time, the location of threat targets is dynamically predicted to generate a dynamic threat situation map. Step 3: Construct a multi-objective route optimization model, define a multi-objective weighted cost function that includes route length cost, altitude cost, energy consumption cost, safety cost and threat cost, and set multiple constraints covering maneuverability, terrain and altitude, range and terminal, swarm coordination and no-fly zones; Step 4: Solve the model based on the improved ant colony algorithm, design and maintain composite environmental pheromones in a differentiated manner, construct multi-angle heuristic information and perform weighted integration, generate an initial feasible route through a feasible route incremental construction method that considers spatiotemporal constraints, and then complete the final route planning through dynamic updates of pheromones.

[0011] Further, in step 1, The initial state set includes the initial release position, initial execution time, flight speed range, and initial flight orientation of the unmanned sub-cluster cluster; The target attribute set includes the target spatial coordinates, the scheduled arrival time, the location of the final task, and the task time window; The waypoint information set includes the spatial coordinates, estimated arrival time, heading angle, pitch angle range, and probability of being detected for any waypoint on the route; The threat set includes air detection threats, traditional threats, new countermeasures, and communication jamming countermeasures; The environmental situation set includes the mission area, terrain elevation, visibility view, electromagnetic environment map, weather threat airspace, and terrain slope information; The constraint set includes mobility constraints, mission terminal constraints, spatiotemporal coordination constraints, and hazardous restricted area constraints.

[0012] Furthermore, in step 2, The terrain data in ASTER GDEM format was downloaded from the geospatial data cloud. The terrain elevation information was indexed using an octree data structure. After the terrain elevation data was completed using the neighborhood interpolation method, the terrain was smoothed to construct the minimum safe flight surface. A visibility view is generated based on the terrain occlusion effect. Points with a value of 1 in the visibility view are in a state of threat visibility, and points with a value of 0 are in the detection blind zone. The initial feasible area of ​​the task is delineated based on the detection blind zone.

[0013] Furthermore, the multi-objective weighted cost function in step 3 specifically includes: The cost of route length is obtained by comparing the actual cumulative length of the route with the straight-line distance from the starting point to the ending point. The altitude cost of a flight path includes absolute altitude cost and relative altitude cost. Absolute altitude cost guides the flight path to prioritize areas with lower altitudes, while relative altitude cost guides the flight path to conform to the terrain. By measuring the degree of change in heading and pitch angle between adjacent waypoints, and combining the unit maneuver energy consumption coefficients of different types of platforms in the horizontal and vertical directions, the route energy consumption cost is obtained. The slope of the local neighborhood of a waypoint is considered to have come at the cost of route safety. The cost of air route threats is the sum of the cost of radar detection threats, traditional threats, and new threats.

[0014] Furthermore, the multiple constraints in step 3 are specifically as follows: Maneuverability constraints include maximum turning angle constraints and maximum pitch angle constraints. The maximum turning angle and maximum pitch angle are both taken as the lower limit of the performance of all platforms participating in the mission, and the angle change and pitch angle change of adjacent track segments are limited to not exceeding the corresponding limit values. Terrain and altitude constraints: including terrain collision constraints and flight altitude constraints. Terrain collision constraints ensure that the waypoint flight altitude exceeds the preset minimum flight altitude, while flight altitude constraints limit the waypoint's altitude relative to the terrain to not exceed the preset maximum flight altitude. Range and terminal constraints: including maximum range constraints and terminal direction constraints. The maximum range constraint ensures that the total route length does not exceed the minimum and maximum range of all platforms. The terminal direction constraint ensures that the error between the terminal bearing and the preset bearing does not exceed the set error tolerance. Cluster coordination constraints include inter-route collision constraints, communication distance constraints, and time coordination constraints. Inter-route collision constraints ensure that the spatial distance between waypoints on different routes is not less than the preset safe distance and the time difference is not less than the safe time interval. Communication distance constraints ensure that the maximum distance between adjacent routes does not exceed the effective communication distance. Time coordination constraints ensure that there is an intersection between the route arrival time and the mission time window. No-fly zone constraints: Ensure that all flight destinations do not fall within objectively restricted areas, areas with extreme weather, or areas with strong electromagnetic interference.

[0015] Furthermore, in step 4, The composite environmental pheromone includes terrain feature pheromone, prohibition pheromone, memory pheromone, and search pheromone; Among them, the terrain feature pheromone is a non-volatile type, with a higher concentration set for low-altitude areas and a minimum concentration set for high-altitude areas; The prohibition pheromone sets the minimum concentration in high-threat areas and no-fly zones, and all sub-clusters share this pheromone list and it does not volatilize; The memory pheromone is updated based on the memory matrix. When the route is successfully constructed, the corresponding value of the memory matrix decreases by a set unit step size. When the route fails, the corresponding value increases by the same unit step size. The concentration of the memory pheromone is negatively correlated with the value of the memory matrix. The initial concentration of the search pheromone is consistent with that of the terrain feature pheromone. Each sub-cluster is maintained independently, and accumulation and volatilization are carried out during the iteration process.

[0016] Furthermore, in step 4, The specific process of the feasible incremental route construction method considering spatiotemporal constraints is as follows: Initialization: Set the route start point, end point, maximum number of extensions, minimum straight distance, and minimum flight altitude. Use a mixed forward and reverse initialization strategy, with the end point as the construction start point in most cases, and use the end direction constraint as the initial extension direction of the candidate waypoints. Candidate waypoint expansion: Based on the current waypoint's heading, candidate waypoints are generated at equal intervals within a horizontally feasible sector at a set angle. The elevation of a candidate waypoint is the terrain elevation at the corresponding location plus the minimum flight altitude. Constraint verification: sequentially verify whether the candidate waypoints are within the mission area and not in the no-fly zone, whether they meet the maneuverability constraints, and whether they meet the spatiotemporal constraints of cluster collaboration. Candidate waypoints that meet all constraints are added to the candidate set. Route generation: If the candidate set is not empty, the next waypoint is selected based on the selection probability determined by the composite environmental pheromone and comprehensive heuristic information. The process is iterated and extended to the destination. After the route is successfully constructed, an interpolation method is used for smoothing.

[0017] A route planning system based on multidimensional information representation and evolutionary computation: The route planning system includes the following modules: task requirements and scenario analysis, multi-dimensional information processing and threat prediction, multi-objective route optimization model construction, and route planning: The task requirements and scenario analysis module clarifies the interaction between terrain obstacles, counter-threat, environmental factors and route planning, defines the requirements for stealth, security and cooperation that the route must meet, and abstracts unmanned swarms, targets, threats and environmental elements into an initial state set, a target attribute set, a waypoint information set, a threat set, an environmental situation set and a constraint set. The multidimensional information processing and threat prediction module is based on terrain data and completes the elevation information through neighborhood interpolation. It smooths the terrain to construct the minimum safe flight surface, determines the detection blind zone based on the terrain occlusion effect and delineates the initial mission feasible area, and dynamically predicts the location of threat targets to generate a dynamic threat situation map. The multi-objective route optimization model construction module defines a multi-objective weighted cost function that includes route length cost, altitude cost, energy consumption cost, safety cost, and threat cost, and sets multiple constraints covering maneuverability, terrain and altitude, range and terminal, swarm coordination, and no-fly zones. The route planning module solves the model based on the improved ant colony algorithm, designs and maintains composite environmental pheromones in a differentiated manner, constructs multi-angle heuristic information and integrates it in a weighted manner, generates an initial feasible route by considering the spatiotemporal constraints of the feasible route incremental construction method, and then completes the final route planning by dynamically updating the pheromones.

[0018] An electronic device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the above method.

[0019] A computer-readable storage medium for storing computer instructions that, when executed by a processor, implement the steps of the above-described method.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention first analyzes the characteristics of flight path missions and the interaction relationships among key scenario elements during low-altitude flight of UAV swarms, and then establishes a hierarchical flight path planning framework for UAV swarms. This framework combines mission area line-of-sight detection, flight path safety, and mission spatiotemporal requirements to construct a flight path planning model based on multi-dimensional information joint representation.

[0021] First, the terrain within the mission space is smoothed to obtain a safe flight surface. Second, the threat detection blind zone created by terrain occlusion is used as the initial range of the mission area, and the mission attributes of potential waypoints in the mission space are analyzed to determine their feasibility in the route construction process. Subsequently, to ensure the low detectability of the resulting route in the future, the movement trend and spatial distribution of potential countermeasures are predicted by combining the terrain slope changes within the mission area. This yields a potential regional threat situation, and a multi-objective route optimization model under multiple constraints is established.

[0022] To solve this model, an initial scheme construction method is proposed to generate feasible routes that meet mission constraints through reverse decoupling, thereby ensuring spatiotemporal coordination constraints among different routes. Based on this, an ant colony search solution scheme based on a composite information structure is proposed. This scheme establishes a differentiated maintenance mechanism to adjust environmental information during the evolution process, promoting the exploration of the mission area by different individuals and effectively preventing routes from entering high-threat areas. Finally, simulation results verify the effectiveness of the proposed route planning solution framework, thus providing a reference for UAV swarms to obtain approach route schemes that meet mission requirements. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the drone swarm collaborative trajectory planning framework of the present invention; Figure 2 A schematic diagram for calculating the threat range considering the terrain shielding effect, including (a) a three-dimensional elevation map of the mission area and (b) a top view of the mission area; Figure 3 This is a schematic diagram of the threat in a local mission area considering motion performance limitations under terrain occlusion effects, where (a) shows the regional situation distribution. N pred =5), (b) Regional situation distribution ( N pred =10); Figure 4 The images show the initialization effects of environmental information obtained by different methods, including (a) uniform initialization and (b) settlement initialization. d c=1、(c) Settlement Initialization d c =0.3; Figure 5 To account for the threat situation map of the mission area under terrain obstruction conditions, (a) the potential range of threat impact and (b) the probability distribution of threat visibility; Figure 6 Initialize the results for relevant information about the task area, including (a) the area slope map and (b) pheromone initialization. d c =0.3; Figure 7 The final collaborative route is shown from different perspectives, including (a) the three-dimensional mission space and (b) the threat situation space. Figure 8 The trends in airway risk include (a) changes in detection threats and (b) changes in interference threats. Figure 9 The curves showing the changes in flight path angles (Flight 1) are shown, where (a) is the heading angle, (b) is the change in heading angle, (c) is the pitch angle, and (d) is the change in pitch angle. Figure 10 The relative flight altitude changes for different routes are shown, where (a) is route one and (b) is route four. Figure 11 The distribution of different types of pheromones during the iteration process includes (a) the search pheromone distribution and (b) the memory pheromone distribution.

[0024] Figure 12 Pseudocode for a feasible route construction algorithm based on incremental expansion; Figure 13 This is the pseudocode for a cluster cooperative route planning algorithm based on multi-level information representation. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Unless otherwise specified, the experimental methods used in the following examples are conventional methods. Unless otherwise specified, the materials, reagents, methods, and instruments used are all conventional materials, reagents, methods, and instruments in the art, and can be obtained commercially by those skilled in the art.

[0027] This invention proposes a hierarchical spatiotemporal cooperative route planning method for low-altitude flight of UAV swarms. It analyzes the interaction relationships between key elements in the mission scenario and constructs a corresponding solution model based on this analysis. The overall process is as follows: Figure 1 As shown: Step 1: Task requirements and scenario analysis, clarify the interaction between environmental elements and route planning, define the core requirements of the route, and abstract scenario elements such as unmanned swarm, target, threat, and environment into six sets: initial state set, target attribute set, waypoint information set, threat set, environmental situation set, and constraint set. First, the relevant mission attributes of different types of elements in the mission scenario are analyzed from multiple dimensions. Based on this, the terrain occlusion effect is comprehensively considered, and the mission threat space is modeled based on a hierarchical representation. Further analysis is conducted on the constraints faced by the unmanned swarm during flight, such as maneuverability, combat mission, and spatiotemporal coordination. Combining threat costs, energy costs, and mission constraint violations, an unmanned swarm trajectory planning model is constructed.

[0028] Assuming that environmental information within the mission area, as well as the prior spatial distribution of targets and threats, has been obtained through reconnaissance before the unmanned swarm executes its mission, the focus is on the impact of route planning on overall spatiotemporal coordination performance when planning the swarm's safe flight path. Furthermore, platforms with the same objective often participate in mission execution as task groups; therefore, during the planning process, the entire unmanned swarm with different mission capabilities can be considered as a set of point masses. Based on this assumption, the trajectory planning problem for unmanned swarms can be described using a multivariate array approach. ,in: This represents unmanned sub-clusters corresponding to objectives with different spatial distributions. The initial state attributes, where This represents its initial release position in the task space. This indicates the initial time of task execution between different clusters, serving as a time reference for arrival at the task area. This indicates the flight speed range of different types of platforms participating in the same mission, used for estimating arrival time. N type The platform category that participates in the task execution. This indicates the initial flight orientation of different clusters, centered on the release point and with the northeast as a reference.

[0029] This represents the task attributes related to the target, where The spatial coordinates of the target can be set based on prior reconnaissance. For the scheduled arrival time of the current target, This indicates the starting position of the mission, centered on the target and with the northeast as a reference. It should be noted that if the unmanned swarm is divided into multiple groups starting from different release points, the corresponding initial attributes need to be set for each group. This enables task-level collaboration. Furthermore, considering the performance limitations of unmanned platforms, there is no situation where transition mission collaboration is possible between large, sparsely distributed targets within the task space. The route must be completed within the pre-defined mission time window.

[0030] This indicates the mission attributes of the unmanned swarm during the route planning process, among which... This represents the spatial coordinates of any waypoint on a single cluster route. Indicates the expected destination from the starting point. j The time required can be approximated as: ,in The method for calculating route length uses the cumulative Euclidean distance between two adjacent points in the embodiment, but it can also be adjusted to the corresponding calculation method according to the actual mission requirements. and These represent the range of heading and pitch angles for that waypoint, respectively. This indicates the probability of the waypoint being detected.

[0031] This represents the set of different types of threats present in the environment. This indicates threats related to air detection. Indicating traditional threats, This indicates new countermeasures. This indicates countermeasures against communication interference; This represents the set of situations within the mission area. The task area is indicated by a grid structure used in this embodiment to store task environment situational information, such as... Figure 2 As shown in (a). = The elevation of any location within the task area can be represented by an octree or similar data structure and obtained through neighborhood interpolation. Potential high-threat areas within the task area are then identified, and a visibility view is constructed based on the resulting threat location distribution. ,in This indicates that the point is under threat of visual contact and requires careful consideration during route planning; conversely, if the point is not under threat, it is considered a threat. This indicates that the point is within the detection blind zone. This represents a map of the electromagnetic environment established based on reconnaissance information. This refers to localized unusable airspace caused by potential weather threats in the natural environment, which also needs to be avoided. Indicates environmental elevation The obtained terrain slope information is used to assist in the subsequent selection of feasible points during the route planning process.

[0032] This represents the set of constraints that the route must satisfy. Where C... d C represents the set of maneuverability constraints between adjacent waypoints. t Indicates task terminal constraints, C c Then, from the perspective of spatiotemporal coordination, the constraints that it needs to comply with are represented, C. f This indicates a pre-defined restricted area for dangerous personnel.

[0033] Step 2: Multidimensional information processing and threat prediction. The terrain data is processed to generate a safe flight surface. The detection blind zone is determined based on the terrain occlusion effect to delineate the initial mission feasible area. At the same time, the location of the threat target is dynamically predicted to generate a dynamic threat situation map. During unmanned platform missions, the location distribution of threat sources may change. Considering that countermeasures are typically located on the ground and are significantly affected by terrain slope, their location distribution over a future period can be predicted based on local spatial terrain characteristics. In a grid-based mission environment, to simplify calculations, the possible flight paths of ground threats at time t are discretized as follows: That is, at each moment, only the direction in the 8-neighborhood of the current coordinate can be selected as the direction of motion for the next moment, and the motion decisions of the previous and next moments are independent of each other.

[0034] Considering that ground threats are often limited by their mobility, especially when the terrain slope exceeds a certain threshold, effective deployment becomes difficult. Therefore, terrain slope information in the environment can be used as a basis for deployment. Further restrictions on its potential course in the next stage are shown in Formula 1. (1) The speed at which the threat moves per unit time is assumed to be randomly generated within a certain range. Therefore, after obtaining the maneuver direction and speed of the ground threat, its position coordinates can be updated according to Formula 2, where [•] indicates rounding the result. N pred To predict the time step, Δt is the unit time, and its corresponding terrain elevation can be expressed as:

[0035] (2) To illustrate the feasibility of the proposed strategy Figure 2(b) selects a 300*300 task area and randomly generates corresponding threats at different locations, and further provides the task area situation under different prediction time steps, as shown below. Figure 3 As shown, the maximum feasible slope is In the diagram, the black dot represents the center of the threat location, the red area represents the line-of-sight detectable area of ​​the predicted threat type, and the blue, yellow, and pink areas represent the effective detection areas of other threats (antenna height 20 meters relative to the ground). It can be seen that detection equipment deployed in low-altitude areas experiences blind spots within its detection range due to obstruction by higher-altitude mountains (taking the lower left area as an example). Furthermore, as the prediction step size increases, the potential effective range of the threat gradually expands, as shown by the blue area near the coordinates (225, 100) on the right. Therefore, in practical applications, different prediction step sizes and target movement speeds can be selected to obtain approximate results of the threat situation in the mission area over a future period.

[0036] Step 3: Construct a multi-objective route optimization model, define a multi-objective cost function that includes length cost, altitude cost, energy cost, safety cost and threat cost, and set multiple constraints covering maneuverability, terrain and altitude, range and terminal, swarm coordination and no-fly zones; To further describe the spatiotemporal characteristics and overall cost of the routes, the cluster route structure in Formula 3 is defined based on the above content. Then the route structure between multiple sub-clusters can be represented as .

[0037] (3) Defining route structure Based on the constraint union C p Based on mission requirements, the cluster route planning problem can be modeled as a multi-objective optimization problem under multiple constraints to minimize mission cost, which can be expressed as Equation 4; where to These represent different types of route costs. The weighting coefficients for each indicator are set, and corresponding parameter weights are set for unmanned platforms with different mission capabilities. This indicates that a mapping relationship has been established between task elements and the resulting flight path; and Let these represent the sets of equality and inequality constraints obtained by repartitioning the constraint set above. The costs and constraint calculation methods are defined below.

[0038] (4) Route length cost Given the limited range of small unmanned platforms, missions should be completed while meeting the maximum route length requirement. Therefore, it is necessary to minimize route deviation and fly directly to the target point. To this end, a normalized route length cost is defined in Formula 5 to measure the impact of route deviation, where... It represents the straight-line distance between the starting point and the ending point.

[0039] (5) Altitude Cost To fully utilize blind spots in route planning, the route altitude cost is further subdivided into absolute altitude cost and relative altitude cost, as shown in Formula 6. For the price of absolute route altitude, For the cost of relative flight path altitude, This indicates the preset minimum flight altitude. and These represent the artificially set altitude ranges relative to the ground surface. Absolute altitude cost guides flight paths to prioritize lower-altitude areas, helping to improve the utilization of blind spots; relative altitude cost, on the other hand, applies to the aircraft's altitude relative to the terrain, helping flight paths to better conform to the terrain and reduce the likelihood of detection.

[0040] (6) Route energy consumption cost Due to the limited endurance of small unmanned platforms, it is necessary to limit the energy consumption caused by changes between adjacent waypoints to avoid unnecessary maneuvers. In real-world scenarios, the energy consumption model of a flight path is nonlinearly coupled with the platform type and the external environment. To simplify the analysis, we approximate the energy consumption using a weighted combination of different proportions and the degree of change in heading between adjacent waypoints, as defined in Equation 7. and These represent the energy consumption per unit of maneuver in the horizontal and vertical directions for different types of platforms, respectively.

[0041] (7) Cost of air route safety The mission's stealth requirements necessitate maintaining a low flight altitude, which to some extent limits its maneuverability and increases the threat of collisions with adjacent natural obstacles. The local safety costs for waypoints are defined here as shown in Equation 8. Where N... h Current waypoint The spatial threat of a local set of neighborhood points is assessed by determining its local slope mean.

[0042] (8) The cost of air route threats Based on the consideration of terrain visibility, the threats posed by various environmental factors to the airway were analyzed and the threat costs of the airway segments were assessed to ensure the overall safety of the airway. The overall cost can be expressed as Formula 9, and the corresponding cost is calculated as follows. For traditional threats, As a new type of threat: (9) Define visibility measure: for any two points P in the task environment i and P j If the line connecting points P and P is not obstructed by terrain, then these two points (P and P) are called points (P and P). i ,P j These constitute a line-of-sight relationship. Its mathematical representation is shown in Formula 10: (10) Radar detects threats Subsequent destructive countermeasures are only triggered after detection measures have detected the target. If the flight path enters the effective range... R max At this time, the radar cross-section of the flight path relative to radar can be reduced. To achieve safe and efficient penetration and close-range attack. Considering that the outer envelope of the unmanned swarm during local flight can be approximated by an ellipsoid and that existing publicly available data shows a relatively low probability of swarm detection under terrain obstruction conditions, to simplify the problem analysis, the centerline of the unmanned swarm's flight path is regarded as a single unmanned platform and used as the basis for calculating the radar detection threat at that waypoint.

[0043] Assume that its overall shape conforms to an ellipsoidal distribution and can be simulated using Equation 11. a , b , c For the corresponding simplified model parameters, and These are the pitch and roll angles relative to the radar along the flight path, respectively.

[0044] (11) After obtaining the estimated radar cross-section of the unmanned platform relative to the radar, the current waypoint is further calculated using Equation 12. The probability of detection of unmanned platforms , in The Euclidean distance between the waypoint and the radar threat. For the first j The three-dimensional coordinates of the radar c 1 and c 2 represents radar parameters, which can be set based on prior reconnaissance information and the countermeasures database.

[0045] (12) Assume that the detection probabilities of the radars capable of detecting the route in the mission environment are independent. When any waypoint... When within the detection range of multiple radars, the overall probability of detection is... It can be calculated using formula 13, where N radar A collection of radars designed to directly detect unmanned platforms.

[0046] (13) Combining formulas 10-13, the overall threat cost of detection measures to the airway can be summarized in formula 14: (14) Route Multi-Constraint Analysis: Further defining the relevant task constraints that the route needs to satisfy, the Lagrange multiplier method is used to add some of the constraints from Equation 4 to the cost function. To ensure that the route can meet the performance requirements of different types of unmanned platforms, the following constraints are first defined.

[0047] Maximum rotation constraint In practical applications, the turning maneuvers of unmanned platforms must meet minimum turning angle constraints, which are related to the platform's flight speed and the angular changes of adjacent flight path segments. Therefore, Formula 19 is used to limit the angular changes of adjacent flight path segments to ensure the feasibility of the overall flight path. (19) in The maximum turning angle between different platforms participating in the mission corresponds to the unit with the worst maneuverability among the unmanned platforms. and This represents the horizontal projection of a continuous path segment.

[0048] Maximum pitch angle constraint During the terrain smoothing preprocessing, a minimum safety surface has been generated by considering the vertical maneuverability of the unmanned platform. The purpose of defining constraints here is to ensure that the smoothed path at discrete points meets the corresponding constraints, which can be expressed as Equation 20: (20) in For the maximum climb angle of different platforms, This represents the distance between adjacent waypoints. It should be noted that routes may be adapted to different types of platforms; therefore, maneuverability constraints should be based on the lower bounds for all platforms.

[0049] Terrain collision constraints To prevent the smoothed flight path from colliding with the terrain, the flight altitude of the corresponding flight path sampling point must exceed a predetermined safety threshold (minimum flight altitude). h min Therefore, the constraints in Equation 21 are defined as follows: (twenty one) Flight altitude constraints When a flight path traverses a special area, in order to reduce the probability of detection and damage, the flight altitude must be limited to a reasonable range. This constraint can be expressed as Equation 22, where... h max Maximum flight altitude: (twenty two) Maximum range constraint Due to the limited mission capabilities of small unmanned aerial vehicle (UAV) platforms, their maximum flight range must meet their maximum available range. Within this range, the flight constraint can be expressed as Equation 23, where L i Defined using Formula 5.

[0050] (twenty three) Track direction constraints Considering that the target may also use terrain to reduce its chance of being detected, the obtained flight path needs to meet certain terminal direction constraints to maximize guidance efficiency. The established flight path direction constraints can be expressed as Formula 24, where... The end point is set by humans. To guide the orientation error tolerance.

[0051] (twenty four) Inter-route collision constraints Simultaneously satisfying spatial and temporal constraints among multiple routes corresponding to different sub-clusters is crucial for ensuring route safety. Specifically, when the... i The first of the routes j The waypoint and the first o The first of the routes p Distance between waypoints Less than the specified space safety distance And the estimated arrival time difference between the two Less than A collision is determined when the number of waypoints is equal between adjacent routes. The collision constraint can be expressed as Equation 25. To ensure that the number of waypoints is the same between adjacent routes and to avoid detection errors due to differences in the number of waypoints, this constraint needs to be checked after smoothing interpolation. The same applies to other constraints. : (25) in The distance between corresponding waypoints on adjacent routes can be further expressed as: (26) in This represents the distance between corresponding waypoints on adjacent routes. The time required for the unmanned swarm to reach the waypoint at its average flight speed can be expressed as: .

[0052] Inter-route communication distance constraints Due to terrain obstruction, the effective communication range between sub-clusters will be significantly shortened. To ensure timely information exchange during the mission, it is assumed that the maximum distance between adjacent routes must not exceed the effective communication distance. d com The specific form of this constraint can be expressed as Formula 27: (27) in This can be further expressed as Formula 28: (28) Route time coordination constraints To ensure mission capability against time-sensitive targets, missions must be executed within a specific timeframe, thus placing higher demands on the overall coordination of their approach routes. Assume the time windows for different sub-clusters are... The expected arrival time range will be selected later. It should intersect with the corresponding time window, which can be further expressed as Formula 29: (29) No-fly zone restrictions : Represented as Formula 30, where Used to determine whether a waypoint is located in a no-fly zone due to objective factors, extreme weather conditions, or strong electromagnetic fields. NFZ .

[0053] (30) Step 4: Solve the model based on the improved ant colony algorithm (CMIR-ACO), design composite environmental pheromones and maintain them differently, construct multi-angle heuristic information and integrate it with weights, generate feasible routes through incremental construction, and dynamically update pheromones to complete route planning.

[0054] A model-based solution method based on incremental guided construction and co-evolution: By observing the discrete route structure used in Equation 3 and the optimization problem P2 defined in Equation 4, it can be seen that the unmanned swarm cooperative route planning problem has multi-objective combinatorial optimization characteristics and belongs to a typical NP-hard problem. Unlike continuous mission environments where the overall path characteristics are described by a finite number of path control points, the shapes of obstacles in discrete environments are usually irregular. Blindly increasing the number of control points will lead to a sharp increase in the dimension of the solution space, thus increasing the difficulty of optimization. Ant colony algorithms have a natural advantage in solving discrete combinatorial optimization problems because they can construct feasible solutions in parallel and optimize the solution space through positive feedback mechanisms. To effectively solve problem P2, a co-evolution of multi-level information representation based on improved antcolony algorithm (CMIR-ACO) method is proposed. This method includes a multi-level information initialization method, multi-angle heuristic information design, a feasible solution construction method based on derivation diffusion, and a pheromone differentiation maintenance mechanism.

[0055] To meet the requirements of low-altitude flight, the terrain features and situational information within the mission area were first analyzed, and environmental pheromones were initialized from different perspectives to guide the feasible solution construction process, thus reflecting the interaction characteristics between different mission areas and route segments. Based on this, a differentiated dynamic pheromone update mechanism was designed, taking into account the differences in mission capabilities of different platform types participating in the mission and the threat characteristics of route segments, highlighting key areas in the search process. Simultaneously, based on mission requirement constraints, a multi-angle heuristic information set was designed to ensure waypoints meet different mission objectives. Secondly, to obtain initial route schemes that satisfy the mission's spatiotemporal constraints, the cooperative route planning problem in a three-dimensional environment was transformed into a two-dimensional planar planning problem based on the construction of a threat surface, effectively reducing the dimensionality of the solution problem. Furthermore, the diffusion derivation concept was introduced, combining maneuverability, mission constraints, and environmental feature information to construct a set of candidate feasible points, and a route was generated using a pseudo-random method. Finally, through iterative optimization and adjustment, the cost of mission route combinations was further reduced.

[0056] A differentiated design method for environmental information structures: Unlike traditional ant colony algorithms that use a single pheromone structure to reflect the interaction between individuals and the environment, in the task context considered in this paper, the flight path faces complex interactions between task requirements and the external environment. Therefore, it is necessary to define multiple pheromone structures to reflect the different attributes of waypoints during the flight path construction process. It is assumed that each unmanned swarm corresponds to a group of ant colony individuals. ,in Nc For each individual ant in the corresponding ant colony, each individual corresponds to a set of alternative flight path structures. The pheromone structure under consideration can be represented as follows: Its specific form is defined as follows: Topographic feature pheromones To achieve the goal of constructing low-altitude air routes, it is necessary to guide the selection of routes to lower-altitude areas through terrain masking strategies. Therefore, initial pheromones are differentiated based on factors such as slope and terrain complexity, and potentially feasible areas in the environment are highlighted using a pheromone settling method based on relative terrain height, as shown in Formula 31.

[0057] (31) in It is a parameter characterizing the decay rate of pheromones as a function of relative altitude. τ max and H min These represent the maximum pheromone concentration and minimum altitude in the mission environment, respectively. To enhance the stealth of the flight path, altitudes higher than [a certain value] are [specified]. H max The region is defined as an infeasible region, and the pheromone level in this region is set to a small constant. τ min >0, to reduce the probability that the area will be explored. Figure 4 Terrain feature pheromone maps obtained using different initialization methods are shown. Figure 4 (a) is a pheromone map generated using a uniform initialization method. The black areas represent forbidden areas with high altitudes, and their pheromones are set to... τ min In comparison, Figure 4 (b) and Figure 4 (c) The pheromone settling initialization method using Formula 31 is illustrated. It can be seen that the pheromones are mainly concentrated in the lower-altitude valleys (white areas), and the method effectively distinguishes the impact of terrain elevation changes on the route exploration area. This method, combined with the characteristics of low-altitude penetration missions (utilizing terrain obstruction), guides the algorithm to focus its search in low-altitude areas by utilizing radar detection blind spots in valleys, thus reducing the search space and improving the route's stealth. Considering that environmental information is usually time-invariant, to maintain the stability of the interaction between different nodes and environmental information during the iterative exploration process, thereby encouraging individuals to prioritize exploring lower-altitude areas, this type of pheromone is set to a non-volatile type in the algorithm.

[0058] Prohibit pheromones This type of pheromone indicates the presence of high-threat interference zones or man-made dangerous airspaces in the environment. During initialization, it is uniformly set to... This improves the safety and stealth of flight routes. It should be noted that different sub-clusters participating in route planning maintain the same list of prohibited pheromones, thus achieving a consistent response to external threats, and these pheromones do not evaporate during the iteration process.

[0059] Memory pheromones These pheromones are used to record the feasible states of individuals from different populations at different locations during the search process. To reduce the probability of the algorithm repeatedly visiting already explored areas, a route memory matrix of the same size as the task environment is first initialized. This matrix is ​​used in candidate routes. When construction fails, the set of subsequent feasible nodes is empty, or the route construction is successful. In unit step length Incremental updates are performed on all nodes along the entire route, i.e. Where successful construction is negative and construction failure is positive. This method aims to prioritize the search for potentially feasible regions by identifying infeasible regions. The magnitude of the local memory pheromone can be represented as... It can be seen that, under ideal conditions, individuals in the ant colony will prioritize exploring subsequent nodes with a higher probability of successful construction, which correspond to lower probability of successful construction. value.

[0060] Search for pheromones This type of pheromone is used to record the cumulative pheromone distribution of different individuals in the task space during the iteration process, consistent with the role of pheromones in the ant colony algorithm. In the early stages of iteration, to guide individuals to prioritize exploring low-altitude areas, its initial value can be expressed as... As iterations proceed, the pheromones accumulate and volatilize accordingly to mark and distinguish key feasible areas in the space. Considering the differences in task requirements and environmental interaction characteristics among different sub-clusters, to avoid mutual interference between search pheromones belonging to different populations during the search process, each population independently maintains its own search pheromone structure.

[0061] Based on the defined composite pheromone structure, Equation 32 is used to integrate it to represent candidate waypoints. Local environmental characteristics of the location. To reduce topographic pheromones. The numerical scale effect of the search pheromone increment is used as the initial search value and added to the current search pheromone when the algorithm's search stalls. In order to reflect the varying degrees of interaction between different flight routes and environmental threats, the search pheromones involved in the dynamic adjustment process were analyzed. A dynamic update and adjustment method was designed, and its specific process is as follows: (32) Local dynamic update: To comprehensively consider the impact of factors such as route length, safety, and arrival time on pheromone increment, this section defines the local pheromone update method in Formula 33 to enhance key route segments. As a local evaporation factor, This indicates the security level of the current air route from different types of threats, which can be calculated with reference to Formula 9. and These represent the corresponding platform types. m The stealth capability and energy consumption coefficient of the ant colony. Different individuals in the ant colony perform pheromone differentiation accumulation operation on the constructed route through Formula 33, thereby extracting local details of the interaction between the route and the environment.

[0062] (33) The local volatility coefficient in Formula 33 can be further defined by Formula 34, where T H =0.5 is the safety threshold. Indicates the calculation of alternative routes The average proportion of medium-threat route segments where the threat probability exceeds the safety threshold. It can be seen that pheromones for alternative routes with high threat proportions have a higher volatility, thus reducing the probability of subsequently choosing that waypoint again.

[0063] (34) Global dynamic update: To accelerate the convergence speed of the ant colony, when the population... All individuals in the group completed the flight route. After construction, pheromone enhancement is performed on the routes corresponding to the current best individual using the global update method defined in Formula 35, where... As the global evaporation factor, Q Where is the pheromone constant. The route cost, which does not include the constraint violation portion, is calculated using Formula 4.

[0064] (35) Local diffusion: To alleviate the situation where excessive concentration of pheromones leads to local optima, the method of information diffusion is adopted to adjust the local pheromones in the neighborhood of the globally optimal individual according to formula 36, ​​where... This represents the locally feasible neighborhood corresponding to the current waypoint.

[0065] (36) Multi-angle heuristic information structure design method: In order to accurately characterize the relationship between the local spatiotemporal characteristics of the candidate waypoint set and the task requirements of individuals in the population during the route construction process, a multi-angle heuristic information structure is defined as shown in Equation 37, and its relevant definitions are as follows: (37) Terrain safety Low-altitude flight missions require the flight path to maintain a certain distance from adjacent static obstacles in the environment, such as mountains, to provide safe space for potential maneuvers. To achieve this, the heuristic information of Equation 38 is defined.

[0066] (38) in Current waypoint The set of feasible neighboring nodes.

[0067] The aim is to maximize the potential threat distance between candidate waypoints and environmental obstacles.

[0068] Smooth flight path To reduce the additional energy loss caused by frequent maneuvers between waypoints, a formula was defined. 39 smoothing heuristics are used to prioritize candidate nodes with smaller elevation undulations in their neighborhood.

[0069] (39) in Used to calculate the relative change in heading of feasible nodes in the candidate node set. This indicates the terrain slope corresponding to the candidate node.

[0070] Threat level To ensure the overall safety of the constructed air route, it is necessary to utilize threat blind spots as much as possible in route planning. Therefore, the threat area distribution map generated by the situation prediction in the previous text is used. Formula 40 is used as a benchmark to measure the threat level of candidate waypoints, where .

[0071] (40) Spatial collaboration To reduce the potential collision safety threat posed by the reuse of the same airspace between air routes, Formula 41 is used to measure the distance between corresponding waypoints on different air routes at adjacent times to determine whether there are any safety hazards. .

[0072] (41) Time coordination To ensure the constructed route meets the time window constraint, the estimated arrival times of candidate nodes need to be estimated. Here, the method of accumulating terrain cross sections is used to calculate the candidate node arrival times. To target T p The distance between them. The core idea is to obtain the corresponding three-dimensional elevation by sampling the line connecting the two points at equal intervals, and then summing the results according to the flight path segments. The estimated arrival time is calculated based on the average speed and straight-line flight path. Based on this, the time-coordinated heuristic information in Formula 42 was designed. The preset time window is... It can be seen that this item reaches its maximum value only when the estimated arrival time is within the specified time window, thus prioritizing candidate waypoints with high time coordination.

[0073] (42) Target distance To ensure that the constructed route eventually reaches its destination, a relative distance heuristic term is defined in Formula 43, where... It is used to calculate the estimated distance between neighboring points and the target point and then normalize the calculation.

[0074] (43) Calculating candidate waypoints Before obtaining the comprehensive heuristic information for candidate waypoints, the heuristic terms in Formula 37 need to be normalized. Based on this, the weighted method in Formula 44 is used to obtain the comprehensive heuristic information for the candidate waypoints. Used to adjust the weighting coefficients of heuristic information under different task requirements.

[0075] (44) Combining formulas 31 to 44, the state transition rule for selecting subsequent waypoints from a single waypoint can be obtained as formula 45, where... and These are the corresponding heuristic weight parameters.

[0076] (45) Feasible incremental route construction method considering spatiotemporal constraints: Given a given mission starting point S p and the finish line T p In such cases, Formula 45 is typically used to select candidate nodes. Select and obtain candidate routes While this process can be repeated to extend to collaborative construction across multiple routes, it neglects the differences in maneuverability between platforms and the spatiotemporal coordination between routes. This makes it difficult to obtain candidate routes that meet mission requirements by adjusting local route characteristics on a large mission scale and in a large solution space. To address this issue, a multi-route construction method considering spatiotemporal coordination is proposed, thereby obtaining candidate routes that meet mission constraints during the initial solution construction stage, thus accelerating the iterative process. Its pseudocode is as follows: Figure 12 The algorithm is shown in Algorithm 3-1.

[0077] To satisfy the route maneuvering constraints in Equations 19 and 20 during the construction process, the node expansion approach based on the SparseA* algorithm is used. The core of its construction process consists of the following steps: Determine candidate waypoints (lines 4-5), using the previous waypoint. and current waypoint Based on the coordinate information, determine the current node. The course. Within the horizontally feasible sector of the current course, intervals are divided equally by angles. M Candidate nodes are calculated, where the radius of the sector depends on the unit expansion step size. l The larger the radius, the more candidate points are considered simultaneously. To avoid the challenges to maneuverability posed by frequent waypoint turns, a minimum straight-line distance between waypoints is introduced. L min .

[0078] Safety assessment (line 6). To ensure that the constructed route does not exceed the mission-defined area and does not intersect with the restricted area, restrictions are placed on the safety of nodes during the construction process and on their exceeding of boundaries.

[0079] Maneuverability assessment (line 8). By calculating whether the angle constraints of Formulas 19 and 20 are satisfied between the extended node and the adjacent waypoint, it can be ensured that the constructed route can guarantee the maneuverability of various platforms.

[0080] Spatiotemporal compatibility assessment (lines 9-13). By calculating the spatiotemporal relationship between the extended node and the constructed route, it is determined whether they satisfy the spatiotemporal compatibility constraints in Equations 25 and 27, thereby reducing invalid searches in subsequent iterative search processes.

[0081] The candidate node combination is completed through the above iterative expansion process. The constructed route is then smoothed using interpolation. It should be noted that if the current route is constructed in reverse, the timestamps of its waypoints need to be synchronized and corrected before using Equation 4 to calculate the overall cost of the constructed route. The pseudocode for the proposed CMIR-ACO process is further provided as follows: Figure 13As shown in Algorithm 3-2, its key steps include: Determine the primary route construction. Equations 3-19 and 3-20 state that the constructed route should meet the maneuverability constraints of different types of platforms using that route. Therefore, during construction, routes with higher maneuverability requirements should be prioritized to meet the lower limit of maneuverability among multiple routes.

[0082] Determine the initial construction state (line 6). To satisfy the end-point direction constraint, a hybrid forward and reverse initialization strategy is adopted during the construction process. That is, the endpoint is set as the construction start point with a higher probability, and the end-point direction constraint is used as the initial expansion direction of the candidate node to perform the expansion operation towards the start point.

[0083] Determine the pheromone update method (line 13). Calculate the pheromone increment for the successfully constructed overall route structure and sub-route structure to fully utilize environmental interaction information during the expansion process.

[0084] Simulation Experiment Analysis: The feasibility and effectiveness of the proposed method in solving the low-altitude flight path planning task for unmanned swarms are verified. The simulation is implemented using MATLAB 2020b on a PC with an AMD Ryzen 9 5950X and 32 GB RAM. First, the processing flow of the 3D task scene is introduced. Based on this, the simulation parameters are set according to the task requirements, and the feasibility of the proposed method is judged based on the obtained flight path status. In subsequent experiments, the coordinate settings are all based on the XOY coordinate system rather than actual latitude and longitude.

[0085] Task Scenario and Parameter Settings: Considering the limited endurance of existing small unmanned platforms, the simulation environment area size is set to 25k * 25k. The task terrain data can be downloaded from the Geospatial Data Cloud (https: / / www.gscloud.cn / home) and processed using the scene element analysis described earlier. In this simulation, 30-meter resolution terrain data files in ASTER GDEM format are randomly selected, where each file consists of a 3601 * 3601 matrix. Figure 2 As shown. Therefore, it is necessary to extract a specified size region from it and then perform downsampling and interpolation to obtain a task elevation data matrix of size 300 * 300.

[0086] To predict the location of different threats in the environment and the distribution of line-of-sight threats over a future period, the method described in Section 3.3.2 was used, and the results are as follows: Figure 5 As shown. Orange, light blue, blue, and yellow represent the radar detection range and link interference range within the area, respectively. From... Figure 5(b) It can be seen that under the existing countermeasure deployment conditions, there are feasible routes in the mission space for our side to carry out low-altitude flight missions.

[0087] The types of unmanned platforms participating in the mission and their related parameter ranges, after configuration, are summarized in Table 1. The settings for unit energy consumption and unit heading change energy consumption range are proportionally set for different platform types. In practical applications, the platform type and parameters can be adjusted. Considering that existing publicly available literature lacks a quantitative description of the stealth capabilities of different platform types under terrain obstruction conditions, the overall stealth capabilities are set based on the proportions of the platform's capability characteristics.

[0088]

[0089] Table 1. Relevant parameters of unmanned platforms

[0090] Table 2. Parameters related to the simulation algorithm Based on some of the parameters in Table 2, we can obtain... Figure 6 The initial environmental pheromone distribution in (a) shows that it reflects the guiding role of the terrain features in the region on the flight path and plays a role in reducing the solution space.

[0091] Route Feasibility Verification: To verify the effectiveness of the proposed method, the task scenario in Table 3 was first defined based on the parameter settings in the environmental information structure differentiation design method. Based on this task requirement, the resulting route is as follows: Figure 7 As shown in the figure. The routes corresponding to different types of platforms are represented by red (Route 1), black (Route 2), light blue (Route 3), and blue (Route 4), respectively.

[0092]

[0093] Table 3 Route Construction Information Figure 7 (a) illustrates the final cooperative route scheme under this mission requirement. It can be seen that the route construction scheme using a hybrid forward and reverse initialization meets the terminal direction constraints in Table 3, thus providing a foundation for subsequent missions. Similarly, it can be seen that different routes traverse threat areas to some extent. To visually represent the safety of different routes, the portions of the corresponding routes with threat probabilities exceeding the safety thresholds in Table 2 are bolded. To more clearly illustrate the interaction between different route segments and threats, the following methods are used... Figure 7(b) The threat situation was visualized in the threat landscape space. Taking the blue route as an example, it traversed different types of threat areas near coordinates (100, 80), but did not introduce an excessively high threat probability. This is partly because the calculation of detection probability is related to the line-of-sight relationship, distance, and attitude of the waypoint relative to the threat, and partly to the potential distribution probability of the threat in space. Figure 7 As can be seen from (b), the visibility probability in the mission area is low, which also means that the risk value of the area is low during the prediction time, and therefore the route has a low threat probability when passing through.

[0094] Figure 8 Showing Figure 7 (a) illustrates the various threats faced by different flight segments. Only threats with non-zero threat values ​​are shown here. It can be seen that detection threats pose a significantly greater risk than communication threats. Figure 8 (b) This is because detection threats have a large effective range, and flight routes inevitably traverse their area of ​​influence. Communication threats, on the other hand, are typically used as phased, localized countermeasures; therefore, their effective range is shorter, and their distance can be maintained within a safe zone by controlling waypoints. Figure 8 (a) Taking the radar threat probability of Route 1 as an example, its threat probability fluctuated significantly within a period of about 15 minutes, and then approached 0 in stages before increasing again. This is because the waypoints in this mission area are within the visual blind zone of detection-type threats, which can be seen from... Figure 7 (a) The route segment highlighted in red has been verified. Figure 8 (b) The probability of communication interference along Route 4 increased significantly around 6 minutes and then decreased rapidly. Figure 7 (a) It can be seen that during this period, the flight path turned into the range of the jamming equipment (yellow area), and then the flight path was quickly adjusted to continue flying toward the target in order to avoid increasing the risk.

[0095] To verify that the constructed route can meet the corresponding maneuver constraints, route 1 is used as an example to show its route-related angle variation curves, as follows: Figure 9 As shown. From Figure 9 (a) and Figure 9 (c) It can be seen that the curves of its heading angle and pitch angle changing with time are reasonable and relatively smooth, and can be further processed and input into the controller as reference instructions. Figure 9 (b) and Figure 9 (d) reflects the trends in the heading and pitch angles of the route, showing that their ranges are reasonable and meet the maneuverability constraints in Table 1. Combined with... Figure 7(a) Red Route 1 shows that it made a heading adjustment near the starting point and then flew toward the target. Before approaching the target, it also made a heading adjustment to meet the terminal direction constraint of 180°. During the flight, it needs to adjust its heading to maintain a low relative flight altitude. Therefore, its heading angle and pitch angle change curves are relatively drastic, which also proves the rationality of the route.

[0096] Figure 10 This demonstrates the changes in flight altitude relative to terrain along different flight routes. Figure 10 (a) is an example, where the preset flight altitude range is marked with a black dashed line. It can be seen that its relative flight altitude remains within a reasonable range throughout the flight. Although there are changes in flight altitude during the flight, it does not collide with the terrain, thus ensuring flight safety and demonstrating good terrain-following ability. Its pitch angle change curve is shown in the figure. Figure 9 (d). This situation can also be observed. Figure 10 (b) This is verified. This also shows that it meets the flight altitude constraints of formulas 21 and 22.

[0097] To prove that the route satisfies the no-entry zone constraint in Equation 30 during the construction process, the search pheromones present in the space after the iterative search will be performed. and memory pheromones The distribution of Figure 11 The explanation is provided below. The red circular area represents a pre-defined restricted area (see...). Figure 11 It can be clearly seen that because a feasibility assessment of the restricted area was incorporated during the node expansion process, the final pheromone distribution did not fall into that area, thus ensuring the safety of the flight path. After multiple iterations, it can be seen that... Figure 11 In (a), the pheromones are concentrated near both sides of the line connecting the start and end points, which can guide the search. Figure 11 (b) shows the memory pheromones of failed or reused waypoints during the construction process. It can be seen that they are more concentrated near the restricted area. This is because the maneuverability of different types of platforms needs to be considered during the node expansion process. When a node expands into the restricted area, the set of available nodes becomes empty, resulting in construction failure and updating of the memory pheromones.

[0098] An electronic device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the above method.

[0099] A computer-readable storage medium for storing computer instructions that, when executed by a processor, implement the steps of the above-described method.

[0100] The above provides a detailed description of the route planning method based on multidimensional information representation and evolutionary computation proposed in this invention, and elucidates the principles and implementation methods of this invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A route planning method based on multidimensional information representation and evolutionary computation, characterized in that: The method specifically includes the following steps: Step 1: Task requirements and scenario analysis, clarify the interaction between terrain obstacles, counter-threat, environmental factors and route planning, define the stealth, security and coordination requirements that the route must meet, and abstract the unmanned swarm, target, threat and environmental elements into an initial state set, a target attribute set, a waypoint information set, a threat set, an environmental situation set and a constraint set. Step 2: Multidimensional information processing and threat prediction. Based on terrain data and by supplementing elevation information through neighborhood interpolation, the terrain is smoothed to construct the minimum safe flight surface. Based on the terrain occlusion effect, the detection blind zone is determined and the initial mission feasible area is delineated. At the same time, the location of threat targets is dynamically predicted to generate a dynamic threat situation map. Step 3: Construct a multi-objective route optimization model, define a multi-objective weighted cost function that includes route length cost, altitude cost, energy consumption cost, safety cost and threat cost, and set multiple constraints covering maneuverability, terrain and altitude, range and terminal, swarm coordination and no-fly zones; Step 4: Solve the model based on the improved ant colony algorithm, design and maintain composite environmental pheromones in a differentiated manner, construct multi-angle heuristic information and perform weighted integration, generate an initial feasible route through a feasible route incremental construction method that considers spatiotemporal constraints, and then complete the final route planning through dynamic updates of pheromones.

2. The method according to claim 1, characterized in that: In step 1, The initial state set includes the initial release position, initial execution time, flight speed range, and initial flight orientation of the unmanned sub-cluster cluster; The target attribute set includes the target spatial coordinates, the scheduled arrival time, the location of the final task, and the task time window; The waypoint information set includes the spatial coordinates, estimated arrival time, heading angle, pitch angle range, and probability of being detected for any waypoint on the route; The threat set includes air detection threats, traditional threats, new countermeasures, and communication jamming countermeasures; The environmental situation set includes the mission area, terrain elevation, visibility view, electromagnetic environment map, weather threat airspace, and terrain slope information; The constraint set includes mobility constraints, mission terminal constraints, spatiotemporal coordination constraints, and hazardous restricted area constraints.

3. The method according to claim 2, characterized in that: In step 2, The terrain data in ASTER GDEM format was downloaded from the geospatial data cloud. The terrain elevation information was indexed using an octree data structure. After the terrain elevation data was completed using the neighborhood interpolation method, the terrain was smoothed to construct the minimum safe flight surface. A visibility view is generated based on the terrain occlusion effect. Points with a value of 1 in the visibility view are in a state of threat visibility, and points with a value of 0 are in the detection blind zone. The initial feasible area of ​​the task is delineated based on the detection blind zone.

4. The method according to claim 3, characterized in that: The multi-objective weighted cost function in step 3 specifically includes: The cost of route length is obtained by comparing the actual cumulative length of the route with the straight-line distance from the starting point to the ending point. The altitude cost of a flight path includes absolute altitude cost and relative altitude cost. Absolute altitude cost guides the flight path to prioritize areas with lower altitudes, while relative altitude cost guides the flight path to conform to the terrain. By measuring the degree of change in heading and pitch angle between adjacent waypoints, and combining the unit maneuver energy consumption coefficients of different types of platforms in the horizontal and vertical directions, the route energy consumption cost is obtained. The gradient of the local neighborhood of a waypoint is used to determine the cost to airway safety. The cost of air route threats is the sum of the cost of radar detection threats, traditional threats, and new threats.

5. The method according to claim 4, characterized in that: The multiple constraints in step 3 are specifically as follows: Maneuverability constraints include maximum turning angle constraints and maximum pitch angle constraints. The maximum turning angle and maximum pitch angle are both taken as the lower limit of the performance of all platforms participating in the mission, and the angle change and pitch angle change of adjacent track segments are limited to not exceeding the corresponding limit values. Terrain and altitude constraints: including terrain collision constraints and flight altitude constraints. Terrain collision constraints ensure that the waypoint flight altitude exceeds the preset minimum flight altitude, while flight altitude constraints limit the waypoint's altitude relative to the terrain to not exceed the preset maximum flight altitude. Range and terminal constraints: including maximum range constraints and terminal direction constraints. The maximum range constraint ensures that the total route length does not exceed the minimum and maximum range of all platforms, and the terminal direction constraint ensures that the error between the terminal bearing and the preset bearing does not exceed the set error tolerance. Cluster coordination constraints include inter-route collision constraints, communication distance constraints, and time coordination constraints. Inter-route collision constraints ensure that the spatial distance between waypoints on different routes is not less than the preset safe distance and the time difference is not less than the safe time interval. Communication distance constraints ensure that the maximum distance between adjacent routes does not exceed the effective communication distance. Time coordination constraints ensure that there is an intersection between the route arrival time and the mission time window. No-fly zone constraints: Ensure that all flight destinations do not fall within objectively restricted areas, areas with extreme weather, or areas with strong electromagnetic interference.

6. The method according to claim 5, characterized in that: In step 4, The composite environmental pheromone includes terrain feature pheromone, prohibition pheromone, memory pheromone, and search pheromone; Among them, the terrain feature pheromone is a non-volatile type, with a higher concentration set for low-altitude areas and a minimum concentration set for high-altitude areas; The prohibition pheromone sets the minimum concentration in high-threat areas and no-fly zones, and all sub-clusters share this pheromone list and it does not volatilize; The memory pheromone is updated based on the memory matrix. When the route is successfully constructed, the corresponding value of the memory matrix decreases by a set unit step size. When the route fails, the corresponding value increases by the same unit step size. The concentration of the memory pheromone is negatively correlated with the value of the memory matrix. The initial concentration of the search pheromone is consistent with that of the terrain feature pheromone. Each sub-cluster is maintained independently, and accumulation and volatilization are carried out during the iteration process.

7. The method according to claim 6, characterized in that: In step 4, The specific process of the feasible incremental route construction method considering spatiotemporal constraints is as follows: Initialization: Set the route start point, end point, maximum number of extensions, minimum straight distance, and minimum flight altitude. Use a mixed forward and reverse initialization strategy, with the end point as the construction start point in most cases, and use the end direction constraint as the initial extension direction of the candidate waypoints. Candidate waypoint expansion: Based on the current waypoint's heading, candidate waypoints are generated at equal intervals within a horizontally feasible sector at a set angle. The elevation of a candidate waypoint is the terrain elevation at the corresponding location plus the minimum flight altitude. Constraint verification: sequentially verify whether the candidate waypoints are within the mission area and not in the no-fly zone, whether they meet the maneuverability constraints, and whether they meet the spatiotemporal constraints of cluster collaboration. Candidate waypoints that meet all constraints are added to the candidate set. Route generation: If the candidate set is not empty, the next waypoint is selected based on the selection probability determined by the composite environmental pheromone and comprehensive heuristic information. The process is iterated and extended to the destination. After the route is successfully constructed, an interpolation method is used for smoothing.

8. A route planning system based on the route planning method of multidimensional information representation and evolutionary computation as described in any one of claims 1 to 7, characterized in that: The route planning system includes the following modules: task requirements and scenario analysis, multi-dimensional information processing and threat prediction, multi-objective route optimization model construction, and route planning: The task requirements and scenario analysis module clarifies the interaction between terrain obstacles, counter-threat, environmental factors and route planning, defines the requirements for stealth, security and cooperation that the route must meet, and abstracts unmanned swarms, targets, threats and environmental elements into an initial state set, a target attribute set, a waypoint information set, a threat set, an environmental situation set and a constraint set. The multidimensional information processing and threat prediction module is based on terrain data and completes the elevation information through neighborhood interpolation. It smooths the terrain to construct the minimum safe flight surface, determines the detection blind zone based on the terrain occlusion effect and delineates the initial mission feasible area, and dynamically predicts the location of threat targets to generate a dynamic threat situation map. The multi-objective route optimization model construction module defines a multi-objective weighted cost function that includes route length cost, altitude cost, energy consumption cost, safety cost, and threat cost, and sets multiple constraints covering maneuverability, terrain and altitude, range and terminal, swarm coordination, and no-fly zones. The route planning module solves the model based on the improved ant colony algorithm, designs and maintains composite environmental pheromones in a differentiated manner, constructs multi-angle heuristic information and integrates it in a weighted manner, generates an initial feasible route by considering the spatiotemporal constraints of the feasible route incremental construction method, and then completes the final route planning by dynamically updating the pheromones.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method of claim 8.

10. A computer-readable storage medium for storing computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method of claim 8.

Images