Unmanned aerial vehicle cluster surveying and mapping task dynamic allocation and path optimization method

Abstract

The invention provides an unmanned aerial vehicle cluster surveying and mapping task dynamic allocation and path optimization method, and relates to the technical field of unmanned aerial vehicle cluster cooperative control, and the method comprises the following steps: S1, carrying out the multi-source data fusion and dynamic task modeling, generating a single unmanned aerial vehicle initial path, and carrying out the multi-unmanned aerial vehicle cooperative adjustment; s2, task allocation based on an improved ant colony algorithm; s3, collaborative path optimization and conflict avoidance; s4, task dynamic redistribution and cluster collaboration; the dynamic adaptability is high, dynamic adjustment of task allocation is achieved through real-time fusion of multi-source data, the response time is shorter than 2 s, and the emergency scene task completion rate is increased by 30%; global optimization is remarkable, the improved ant colony algorithm is combined with collaborative path planning, the total flight distance of the cluster is reduced by 25%, and the surveying and mapping efficiency is improved by 40%; robustness is prominent, a dynamic redistribution mechanism effectively processes unmanned aerial vehicle faults, environment sudden changes and other abnormalities, and the task redistribution success rate is larger than 95%. The precision is guaranteed, and the terrain complexity is matched with the load type, so that the percent of pass of surveying and mapping data is improved from 82% to 96%.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle cluster cooperative control, in particular to a kind of unmanned aerial vehicle cluster surveying and mapping task dynamic allocation and path optimization method, it is applicable to the large-area surveying and mapping operation under complex terrain environment, can realize the intelligent allocation of task and the global optimization of path, improves surveying and mapping efficiency and accuracy. BACKGROUND

[0002] Unmanned aerial vehicle cluster surveying and mapping realizes rapid large-area data acquisition by multi-machine cooperation, and has been widely applied in geographic information surveying and mapping, disaster emergency monitoring and other fields.

[0003] The prior art has the following bottlenecks: Task allocation is static: traditional methods divide tasks based on pre-set regions (such as grid partitioning), which cannot respond to environmental changes (such as sudden obstacles and unmanned aerial vehicle failures) in real time, resulting in task repetition or coverage blind spots; Path planning is isolated: single unmanned aerial vehicle path optimization does not consider cluster cooperation constraints (such as communication links and endurance balance), often resulting in local optimization but low global efficiency; Poor dynamic adaptability: lack of real-time state sensing and feedback of unmanned aerial vehicles (power, payload, communication quality), which can easily cause cluster conflicts during task re-allocation.

[0004] In the prior art, the task allocation method proposed in patent CN113238412A is only based on static task priority and does not integrate real-time environmental data; the traditional A* algorithm is used in the document "Multi-UAV Cooperative Path Planning Algorithm", which does not solve the coupled optimization problem of multi-machine collision avoidance and endurance balance. Therefore, there is an urgent need for a systematic method that integrates dynamic task modeling, intelligent allocation algorithm and cooperative path optimization to achieve efficient surveying and mapping operations in complex scenarios. SUMMARY

[0005] The present application provides a method for dynamic allocation and path optimization of unmanned aerial vehicle cluster surveying and mapping tasks, which solves the problems of static allocation, isolated planning and poor adaptability of traditional methods through multi-source data fusion modeling, improved ant colony algorithm and dynamic re-allocation mechanism, and realizes the real-time of task allocation, the global optimality of path planning and the robustness of cluster cooperation.

[0006] The present application provides a method for dynamic allocation and path optimization of unmanned aerial vehicle cluster surveying and mapping tasks, which specifically includes the following steps: A method for dynamic allocation and path optimization of unmanned aerial vehicle cluster surveying and mapping tasks includes the following steps: S1 Multi-source data fusion and dynamic task modeling: Environmental perception: The drone is equipped with a visible light camera, lidar, and RTK-GNSS module to collect real-time terrain elevation data (accuracy ≤ 5cm), obstacle coordinates (3D point cloud density ≥ 10 points / m 2 ) and meteorological parameters (wind speed, humidity); the ground station obtains a priori maps of large areas through satellite remote sensing with a resolution of ≤0.2m.

[0007] Task decomposition: Divide the survey area into irregular sub-areas (based on Delaunay triangulation), and define task attributes for each sub-area: area Sᵢ , terrain complexity Cᵢ (calculated by elevation standard deviation), time window Tᵢ (such as the no-fly period against sunlight).

[0008] Drone state modeling: Constructing the real-time state vector U of the drone j =[ Ej , Lj , Cj ,P j ],in Ej is the remaining power, Lj is the payload type (orthophoto / oblique photography), Cj is the communication quality, P j is the current location coordinate.

[0009] Furthermore, the terrain elevation data acquisition accuracy in S1 is ≤5cm, and the drone state vector includes the remaining power, payload type, communication quality and location coordinates.

[0010] S2 task allocation based on improved ant colony algorithm: Constructing task allocation graph: Taking drone node U j and task nodes Tᵢ is a vertex, and the edge weight includes the task execution cost ,in is the cruising speed of the drone, is the distance from the current position to the starting point of the task, and α and β are weight coefficients.

[0011] Pheromone renewal strategy: Introducing dynamic evaporation rate ( is the initial evaporation rate, is the environmental change sensitivity coefficient). After the task is completed, press Update pheromones and give priority to cost-effective tasks.

[0012] Taboo table optimization: Combined with UAV payload constraints (e.g., oblique photography UAVs are prioritized for complex terrain missions) and endurance constraints (remaining battery power must meet the round-trip mission requirements), the taboo table access conditions are dynamically adjusted to avoid invalid task assignments.

[0013] Further, the task execution cost function in S2 includes task area, terrain complexity, UAV speed, distance, and power factors, the evaporation rate of pheromone is dynamically adjusted according to the environment, and steps S1-S4 are cooperatively operated to realize a task allocation response time < 2s, a total flight distance of the cluster is reduced by 20%-30%, and a sudden scene task completion rate is ≥95%.

[0014] S3 cooperative path optimization and conflict avoidance: Global path initial planning: for the allocated sub-tasks, an improved RRT* algorithm is used to generate an initial path for a single UAV, and the objective function is where L is the path length, is the obstacle avoidance cost, is the endurance balancing weight.

[0015] Multi-machine cooperative optimization: a cluster conflict detection matrix is constructed, when the distance between the paths of two machines is less than a safety threshold (50m), the local path point is adjusted by a virtual potential field method, and the potential field force formula is , (k is the repulsive force coefficient, d is the current distance, and d0 is the safety distance).

[0016] Dynamic re-planning trigger: when the UAV power is lower than a threshold (30%), the load fails, or a sudden obstacle (detection distance < 200m) occurs, local re-planning is triggered, and a safe path within 10 minutes in the future is generated by a rolling horizon optimization (MPC).

[0017] Further, the cooperative path optimization objective function in S3 includes path length, obstacle avoidance cost, and endurance balancing weight, and the conflict avoidance uses a virtual potential field method to calculate repulsive force to adjust the path.

[0018] S4 task dynamic re-distribution and cluster cooperation: Real-time state monitoring: the ground station receives UAV state data at a frequency of 50Hz, and constructs a task completion progress matrix , is the UAV j execution progress of task i at time t.

[0019] Redistribution condition judgment: when a task timeout (completion progress < 50% and remaining time < T i / 2), UAV failure (communication interruption > 10s), or new task insertion (such as emergency monitoring area) occurs, the dynamic re-distribution process is started.

[0020] Redistribution algorithm: a multi-objective optimization algorithm (NSGA-II) based on Pareto optimality is used, the optimization objectives are minimization of task completion time, balancing of UAV load, and minimization of energy consumption cost, a new task allocation scheme is generated and broadcast to the cluster.

[0021] Further, the dynamic re-distribution trigger condition in S4 includes task timeout, UAV failure or new task insertion, and the re-distribution algorithm adopts NSGA-II to realize multi-objective optimization.

[0022] The present application provides a UAV cluster mapping task dynamic distribution and path optimization method, which has the following beneficial effects: Strong dynamic adaptability: through real-time fusion of multi-source data, dynamic adjustment of task distribution is realized, response time < 2s, and sudden scene task completion rate is improved by 30%; Global optimization is significant: improved ant colony algorithm combined with cooperative path planning makes the total flight distance of the cluster reduced by 25%, and the mapping efficiency is improved by 40%; Robustness is outstanding: the dynamic re-distribution mechanism effectively handles abnormal situations such as UAV failure and environmental mutation, and the task re-distribution success rate is > 95%; Accuracy guarantee: the terrain complexity and load type matching strategy makes the mapping data qualified rate improved from 82% to 96%. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments will be briefly introduced below.

[0024] The drawings described in the following description only relate to some embodiments of the present application, and are not limited to the present application.

[0025] In the drawings: Figure 1 The present application method flow chart is shown; Figure 2 The present application core module method schematic diagram is shown. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments. Based on the described embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0027] Embodiment: please refer to Figures 1 to 2 : The application provides a UAV cluster surveying and mapping task dynamic allocation and path optimization method, which specifically comprises the following steps: S1 multi-source data fusion and dynamic task modeling; environment perception: the UAV is equipped with a visible light camera, a laser radar, and an RTK-GNSS module, and terrain elevation data (an accuracy of ≤5 cm), obstacle coordinates (a three-dimensional point cloud density of ≥10 points / m 2 ) and meteorological parameters (wind speed, humidity) of a surveying and mapping area are collected in real time; a ground station obtains a prior map of a large area through satellite remote sensing, and the resolution is ≤0.2 m.

[0028] Task decomposition: the surveying and mapping area is divided into irregular sub-areas (based on Delaunay triangulation), and the task attributes of each sub-area are defined: an area Sᵢ , terrain complexity Cᵢ (computed by an elevation standard deviation), a time window Tᵢ (such as a no-fly period in backlight).

[0029] UAV state modeling: a real-time state vector U j of the UAV is constructed Ej , Lj , Cj , P j , wherein Ej is the remaining power, Lj is the load type (orthographic camera / tilt photography), Cj is the communication quality, and P j is the current position coordinates. In S1, the terrain elevation data collection accuracy is ≤5 cm, and the UAV state vector comprises the remaining power, the load type, the communication quality and the position coordinates.

[0030] S2 task allocation based on an improved ant colony algorithm; a task allocation graph is constructed: the UAV node U j and the task node T ᵢ are vertices, and the edge weight value comprises a task execution cost , wherein is the UAV cruising speed, is the distance from the current position to the task starting point, and α and β are weight coefficients.

[0031] pheromone update strategy: a dynamic evaporation rate is introduced, is an initial evaporation rate, is an environmental change sensitivity coefficient), the pheromone is updated according to after the task is completed, and a high cost-effective task edge is preferentially selected.

[0032] Tabu list optimization: combined with UAV payload constraints (e.g. tilt-rotor UAVs prefer to assign complex terrain tasks) and endurance constraints (remaining battery power needs to meet the task round-trip demand), dynamically adjust the tabu list admission conditions to avoid invalid task assignments. The task execution cost function in S2 contains task area, terrain complexity, UAV speed, distance, and battery power factors. The pheromone evaporation rate is dynamically adjusted according to environmental changes. Steps S1-S4 run cooperatively to achieve a task assignment response time < 2s, a 20-30% reduction in total cluster flight distance, and a 95% task completion rate in emergency scenarios.

[0033] S3 cooperative path optimization and conflict avoidance; global path initial planning: for the assigned subtasks, an improved RRT* algorithm is used to generate the initial path for a single UAV, with the objective function being where L is the path length, is the obstacle avoidance cost, is the endurance balancing weight.

[0034] Multi-UAV cooperative optimization: a cluster conflict detection matrix is constructed, and when the distance between the paths of two UAVs is less than the safety threshold (50m), the local path points are adjusted using the virtual potential field method, with the potential field force formula being (k is the repulsive force coefficient, d is the current distance, and d0 is the safety distance).

[0035] Dynamic re-planning trigger: when the UAV's battery level is below the threshold (30%), the payload fails, or a sudden obstacle is detected (detection distance < 200m), local re-planning is triggered, and a safe path within the next 10 minutes is generated using the moving horizon optimization (MPC). The cooperative path optimization objective function in S3 includes path length, obstacle avoidance cost, and endurance balancing weight, and the conflict avoidance uses the virtual potential field method to calculate the repulsive force to adjust the path.

[0036] S4 task dynamic reassignment and cluster cooperation; real-time state monitoring: the ground station receives UAV state data at a frequency of 50Hz, and constructs a task completion progress matrix , is the UAV j's execution progress for task i at time t.

[0037] Reassignment condition determination: when a task timeout occurs (completion progress < 50% and remaining time < T i / 2), a UAV failure (communication interruption > 10s), or a new task is inserted (e.g. emergency monitoring area), the dynamic reassignment process is started.

[0038] Redistribution algorithm: A multi-objective optimization algorithm based on Pareto optimality (NSGA-II) is used to minimize task completion time, UAV load balancing, and energy cost. A new task allocation scheme is generated and broadcast to the cluster. The dynamic redistribution trigger conditions in S4 include task timeout, UAV failure or new task insertion. The redistribution algorithm uses NSGA-II to achieve multi-objective optimization.

[0039] In the embodiments of the present disclosure, a 100km 2 Take the mapping project as an example (deploy 5 DJI Matrice300RTK drones): Data collection: The drone swarm obtains centimeter-level positioning data through RTK-GNSS, and generates a DEM model using a laser radar to identify 3 steep terrains (Ci>0.8) and 2 temporary no-fly zones; Task allocation: The improved ant colony algorithm allocates complex terrain tasks to drones with tilt photography payloads. The initial allocation scheme reduces the total flight distance by 18km compared to the traditional grid method; Path planning: The RRT* algorithm generates paths that avoid no-fly zones, and the maximum distance between each drone after collaborative optimization is maintained at more than 80m, with no conflict risk; Dynamic adjustment: When the power of No.1 drone decreases to 25%, the ground station triggers redistribution, transferring the remaining tasks to adjacent drones, and the total task completion time only increases by 5 minutes.

[0040] In this paper, the following points need attention: 1. The drawings of the embodiments of the present disclosure only involve the structures involved in the embodiments of the present disclosure, and other structures can refer to the usual design.

[0041] 2. In the case of no conflict, the embodiments of the present disclosure and the features in the embodiments can be combined to obtain new embodiments.

[0042] The above is only a specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present disclosure, which should be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A method for dynamic allocation and path optimization of UAV cluster mapping tasks, characterized by: The following steps are involved: S1 Multi-source Data Fusion and Dynamic Mission Modeling: Collect environmental data and UAV status to build a dynamic model that includes terrain, mission, and UAV status; S2 Task allocation based on improved ant colony algorithm: build task allocation graph and realize intelligent task allocation through dynamic pheromone update and taboo table optimization; S3 Collaborative Path Optimization and Conflict Avoidance: Generates a single drone's initial path and performs multi-drone collaborative adjustments to avoid obstacles and flight conflicts; S4 task dynamic reallocation and cluster collaboration: real-time monitoring of task progress and drone status, triggering the reallocation mechanism to handle abnormal situations.

2. The method for dynamic allocation and path optimization of UAV cluster mapping tasks according to claim 1 is characterized in that: The terrain elevation data acquisition accuracy in S1 is ≤5cm, and the drone state vector includes the remaining power, payload type, communication quality and location coordinates.

3. The method for dynamic allocation and path optimization of UAV cluster mapping tasks according to claim 1 is characterized in that: The task execution cost function in S2 includes the task area, terrain complexity, drone speed, distance and power factors, and the pheromone evaporation rate is dynamically adjusted with environmental changes.

4. The method for dynamic allocation and path optimization of UAV cluster mapping tasks according to claim 1 is characterized in that: The collaborative path optimization objective function in S3 includes path length, obstacle avoidance cost and endurance balance weight. The conflict avoidance adopts the virtual potential field method to calculate the repulsive force to adjust the path.

5. The method for dynamic allocation and path optimization of UAV cluster mapping tasks according to claim 1 is characterized in that: The dynamic reallocation triggering conditions in S4 include task timeout, UAV failure or new task insertion, and the reallocation algorithm uses NSGA-II to achieve multi-objective optimization.

6. The method for dynamic allocation and path optimization of UAV cluster mapping tasks according to any one of claims 1 to 5, characterized in that: The steps S1-S4 are run in coordination to achieve a task allocation response time of less than 2s, a reduction in the total flight distance of the cluster by 20%-30%, and a task completion rate of ≥95% in emergency scenarios.

Images