Unmanned aerial vehicle pre-flight intelligent scheduling method and system based on fixed air route

Abstract

The invention provides an unmanned aerial vehicle pre-flight intelligent scheduling method based on a fixed air route, which comprises the following steps: S1, constructing an airspace air route topological structure based on air route nodes and road section attributes, and forming an air route relationship network; s2, according to the air route relation network and a flight plan submitted by a user, obtaining air routes, nodes and intersections required by the flight plan; s3, predicting flight conflicts and dynamically adjusting the take-off time to ensure that no time conflict exists between the starting point and the ending point on the flight path and meet the minimum safety interval requirement; and S4, scheduling the flight tasks, so that all the flight tasks are completed in the applied time window, and the minimum take-off interval requirement is met. The invention further provides an intelligent scheduling system for the unmanned aerial vehicle before flight based on the fixed air route. The system comprises a road network layer, a planning layer, a decision-making layer and a scheduling layer. According to the invention, the efficiency and accuracy of front-flight shift are improved, the flight conflict can be effectively predicted and solved, and the route resource allocation is optimized.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicles, and in particular relates to a method and system for intelligent pre-flight scheduling of unmanned aerial vehicles. Background Art

[0002] With the rapid development of the low-altitude economy, airway resources are becoming increasingly scarce, and the problem of scheduling fixed-route flights within strips of airspace is becoming increasingly prominent. Traditional flight scheduling methods rely primarily on manual experience, which is inefficient and unable to meet the growing demand for flights. Manual scheduling is not only time-consuming and labor-intensive, but also prone to human error, leading to flight conflicts and delays, and failing to ensure flight safety and the efficient use of airspace resources. In complex strips of airspace, with a large number of aircraft and complex flight routes, traditional methods are unable to accurately predict and resolve flight conflicts, making it difficult to achieve the rational planning and scheduling of flight missions. Summary of the Invention

[0003] To this end, the technical problem to be solved by the present invention is to provide a method and system for intelligent pre-flight scheduling of drones based on fixed routes, which can reasonably plan and schedule drone flight missions and optimize route resource allocation.

[0004] In a first aspect, the present invention provides a method for intelligent pre-flight scheduling of UAVs based on a fixed route, comprising:

[0005] Step S1, constructing an airspace route topology structure based on route nodes and route segment attributes to form a route relationship network;

[0006] Step S2, obtaining the routes, nodes and intersection points required for the flight plan based on the route relationship network and the flight plan submitted by the user;

[0007] Step S3: predict flight conflicts and dynamically adjust takeoff time to ensure that there is no time conflict between the start and end points on the flight path and that the minimum safety interval requirement is met;

[0008] Step S4: Schedule the flight missions so that all flight missions are completed within the requested time window and meet the minimum takeoff interval requirements.

[0009] Furthermore, in step S1, the route relationship network generation step is:

[0010] Get the UTM longitude and latitude of the route;

[0011] Convert the route UTM longitude and latitude into two-dimensional coordinates;

[0012] Obtain information on main roads, branch roads and intersections;

[0013] Calculate the length of each route;

[0014] Establish inter-route linkages.

[0015] Furthermore, in step S2, a depth-first search is used to obtain all possible flight paths, and routes, nodes, and intersection points required for the flight plan are obtained from all possible flight paths.

[0016] Furthermore, step S3 specifically includes:

[0017] Get the time period occupied by the intersection point in the route relationship network;

[0018] Get the starting and ending points of each path;

[0019] Calculate the theoretical time to reach each intersection based on the flight speed and path length, and generate a time window for reaching each intersection based on the theoretical time;

[0020] Check whether the time window of each intersection conflicts with the occupied time period; if there is a conflict, postpone the departure time; if there is no conflict, return the adjusted time and path.

[0021] Furthermore, in step S4, flight mission scheduling includes:

[0022] A flight mission scheduling model is established, and the model objective function is:

[0023] Fitness=Total Priority-α*Time Range-β*Path Diversity

[0024] Among them, Fitness is the fitness, Total Prioritory is the sum of all priorities assigned to the UAV, Time Range is the duration of the flight mission, Path Diversity is the difference in flight missions between routes, and α and β are weight coefficients.

[0025] Furthermore, a genetic algorithm is used to find the optimal solution for the flight mission scheduling model, including:

[0026] Individuals are represented as routes and schedules for all flight missions;

[0027] The population is initialized to randomly assign routes and schedule takeoff times in sequence;

[0028] The selection operation uses tournament selection;

[0029] The crossover operation is a two-point crossover with adjustments to the timing of offspring;

[0030] The mutation operation is to randomly change the flight route allocation of the flight mission and adjust the time;

[0031] Fitness evaluation uses the objective function.

[0032] Furthermore, in step S1, a segmented structure and a hierarchical capacity detection strategy are used to establish a route relationship network.

[0033] Furthermore, in step S1, based on historical UAV flight trajectory data, the time period occupied by the intersection point is predicted through machine learning.

[0034] In a second aspect, the present invention provides a pre-flight intelligent scheduling system for UAVs based on fixed routes, the system comprising:

[0035] Road network layer, generating route relationship network based on route node and segment attributes;

[0036] The planning layer obtains the routes, nodes, and intersections required for the flight plan based on the route relationship network and the flight plan submitted by the user;

[0037] The decision-making layer predicts flight conflicts and dynamically adjusts takeoff times to ensure that there are no time conflicts between the start and end points of the flight path and that minimum safety interval requirements are met.

[0038] The scheduling layer schedules flight missions so that all flight missions are completed within the requested time window and meet the minimum take-off interval requirements.

[0039] Furthermore, in the road network layer, the steps for generating the route relationship network are as follows:

[0040] Get the UTM longitude and latitude of the route;

[0041] Convert the route UTM longitude and latitude into two-dimensional coordinates;

[0042] Obtain information on main roads, branch roads and intersections;

[0043] Calculate the length of each route;

[0044] Establishing inter-route linkages;

[0045] In the planning layer, a depth-first search is used to obtain all possible flight paths, and the routes, nodes, and intersections required for the flight plan are obtained from all possible flight paths;

[0046] At the decision-making level, predicting flight conflicts and dynamically adjusting takeoff times specifically include:

[0047] Get the time period occupied by the intersection point in the route relationship network;

[0048] Get the starting and ending points of each path;

[0049] Calculate the theoretical time to reach each intersection based on the flight speed and path length, and generate a time window for reaching each intersection based on the theoretical time;

[0050] Check whether the time window of each intersection point conflicts with the occupied time period; if there is a conflict, postpone the departure time; if there is no conflict, return the adjusted time and path;

[0051] In the scheduling layer, flight mission scheduling includes:

[0052] A flight mission scheduling model is established, and the model objective function is:

[0053] Fitness=Total Priority-α*Time Range-β*Path Diversity

[0054] Among them, Fitness is the fitness, Total Prioritory is the sum of all priorities assigned to the UAV, Time Range is the duration of the flight mission, Path Diversity is the difference in flight missions between routes, and α and β are weight coefficients.

[0055] Beneficial effects:

[0056] The method and system for intelligent pre-flight scheduling of UAVs based on fixed routes provided by the present invention realize a closed loop of the entire process from basic route network generation to intelligent decision-making through a four-layer collaborative architecture, thereby improving the automation and intelligence level of pre-flight scheduling, reducing manual intervention, and lowering the risk of human error.

[0057] The road network layer uses a rule engine and optimization algorithm to monitor flight flow in real time, dynamically adjust route allocation, and predict congested nodes, effectively improving airspace utilization and ensuring smooth routes.

[0058] The planning layer uses depth-first search combined with specific strategies to quickly obtain efficient and fast flight routes to meet the needs of flight plans.

[0059] The decision-makers ensured that there were no time conflicts at key intersections on the flight path through conflict prediction and take-off time adjustment, met the minimum safety interval requirements, and improved flight safety.

[0060] The genetic algorithm of the scheduling layer can optimize the flight mission scheduling under multiple constraints, maximize the total priority, balance the channel task allocation, and improve resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to make the contents of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments of the present invention in conjunction with the accompanying drawings.

[0062] Figure 1This is a flow chart of the intelligent pre-flight scheduling method for drones according to Example 1 of the present invention;

[0063] Figure 2 This is a flow chart of route planning according to embodiment 1 of the present invention;

[0064] Figure 3 This is an architecture diagram of the intelligent pre-flight scheduling system for drones according to Example 2 of the present invention. DETAILED DESCRIPTION

[0065] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. The principles and features of the present invention will be described below in conjunction with the accompanying drawings. It should be noted that the embodiments and features of the embodiments in this application may be combined with each other unless there is a conflict. The embodiments are provided only to illustrate the present invention and are not intended to limit the scope of the invention.

[0066] Example 1

[0067] The intelligent route scheduling method for drones before flight provided in this embodiment realizes the route and time planning of drones by constructing a road network, planning paths, making decision jumps, and scheduling.

[0068] During pre-flight scheduling, the airspace network topology is first constructed to provide foundational data for subsequent layers. Based on this network data, a depth-first search algorithm combined with specific strategies is used to obtain the route information required for flight planning. Based on this information from the planned paths, conflicts are predicted and takeoff times are adjusted. Finally, a genetic algorithm is used to schedule flight missions while satisfying various constraints, maximizing the sum of priorities and optimizing resource allocation.

[0069] like Figure 1 As shown, the steps of this embodiment include:

[0070] 1. Construct an airspace road network topology, establishing connections between road nodes (starting and ending points) and road segment attributes (such as length). Dynamically model and manage the route network using a segmented structure and hierarchical capacity detection strategies. A rules engine monitors segment traffic in real time and executes predefined conflict avoidance rules. Optimization algorithms dynamically adjust route allocation to maximize airspace utilization. Machine learning predicts congestion nodes based on historical data to optimize long-term route planning.

[0071] 2. Considering multiple constraints, a depth-first search (DFS) combined with specific strategies is used to explore potential paths, obtaining the routes, nodes, and intersections required for flight planning, thereby providing efficient and effective routes. Depth-first search begins by selecting a starting point and then retracing when no further progress is possible until all connected vertices have been traversed. A stack is used to store access path information.

[0072] 3. Predict flight conflicts and dynamically adjust aircraft takeoff times to ensure there are no time conflicts at key intersections on the flight path and to meet minimum safety separation requirements.

[0073] 4. Use genetic algorithms to solve the flight mission scheduling problem at multi-channel airports.

[0074] Exemplarily, the steps for generating a route relationship are:

[0075] S11: Obtain the longitude and latitude of the route UTM (Universal Transverse Mercator coordinate system);

[0076] S12: Convert the route UTM longitude and latitude into two-dimensional coordinates;

[0077] S13: Instantiate main roads, branch roads, and intersections based on the road network design;

[0078] S14: Calculate the length of various routes;

[0079] S15: Establish inherent connections between routes;

[0080] S16: Establishment completed.

[0081] Preferably, a depth-first search (DFS) is used to explore potential paths through the constructed route relationship network and the flight plan submitted by the user.

[0082] Obtain route information for the entire flight plan, including routes, nodes, and intersections.

[0083] like Figure 2 As shown, an exemplary specific step is:

[0084] S21: Path Planning

[0085] Get all possible flight paths;

[0086] If there is no valid path, the process will terminate with an error.

[0087] S22: Conflict Detection Preparation

[0088] Extract the occupancy time periods of the intersection points of the existing route network;

[0089] Intersection occupancy periods can be predicted using machine learning when constructing airspace network topology, based on historical drone flight trajectory data. Predicting intersection occupancy periods takes historical drone flight trajectories into account, enabling efficient use of airspace resources.

[0090] Extract key intersection points (the starting and ending points of each road) for each path.

[0091] S23: Timetable calculation

[0092] Calculate the theoretical time to reach each intersection based on the aircraft speed and path length;

[0093] Generate a time window ([arrival time - min_interval, arrival time + min_interval]).

[0094] S24: Conflict Detection and Adjustment

[0095] Check whether the time window of each intersection overlaps with the existing occupancy period;

[0096] If there is a conflict, try to postpone the takeoff time, up to max_attempts times;

[0097] If there is no conflict on a path, the adjusted time and path are immediately returned.

[0098] S25: Termination Condition

[0099] Find a conflict-free path → Return the adjusted time and use the airways, nodes, and intersections in the path as the airways, nodes, and intersections required for the flight plan.

[0100] All paths conflict → TimeoutError is thrown.

[0101] For example, the genetic algorithm is used to arrange the path and time point of each aircraft. The implementation steps are as follows:

[0102] S31: Problem modeling: Represent each flight mission as an object containing attributes such as channel, take-off time, start point, and end point. Consider the resource allocation of two channels (Channel_1 and Channel_2) and handle the different constraints of the main and branch flight missions.

[0103] S32: Conflict resolution mechanism: Check and resolve time interval conflicts of flight missions in the same channel, especially deal with conflicts of branch flight missions at intersection points, and ensure that the minimum safety interval (MIN_INTERVAL) is met by adjusting the take-off time.

[0104] Objective function design: maximize the sum of the priorities of all flight tasks, minimize the time distribution range (the time span for completing all tasks), balance the task allocation of the two channels (reduce the difference in the number of tasks between channels), and form a comprehensive evaluation index by weighted combination of these factors. The fitness function is:

[0105]

[0106] in:

[0107] Fitnness means the fitness value is used as an evaluation index,

[0108] TotalPrioritory represents the sum of priorities used to assign priority routes to UAVs.

[0109] Time Range=0.1(max(Time i )-min(Time i )) represents the task duration, which is used to optimize scheduling efficiency;

[0110] Path Diversity=0.5|Count(First)-Count(Second)| represents the constraint on path utilization.

[0111] The fitness function is the optimization goal of this scheme, and the algorithm seeks the optimal state under multiple constraints.

[0112] S33: Genetic Algorithm Implementation:

[0113] Individual representation: a complete scheduling plan (channels and time arrangements for all flight missions);

[0114] Population initialization: randomly assign channels and schedule takeoff times in sequence;

[0115] Select Action: Use tournament selection;

[0116] Crossover operation: crossover between two points and adjust the timing of offspring;

[0117] Mutation operation: randomly change the channel allocation of tasks and adjust the time;

[0118] Fitness evaluation: Use the above objective function.

[0119] S34: Constraint processing: Ensure that all flight missions are completed within the requested time window, meet the minimum takeoff interval requirements, and handle special constraints of branch missions at intersection points.

[0120] Example 2

[0121] This embodiment is a pre-flight intelligent scheduling system for drones based on fixed routes. The system architecture is: adopting a four-layer collaborative architecture of "route network-planning-decision-scheduling". By building a road network, planning routes, making decisions and finally scheduling, the route and time of drones are planned. Figure 3 .

[0122] Road network layer: Construct the airspace road network topology, establish connection relationships based on road nodes (starting points, end points) and road segment attributes (such as length), and use segmented structures and hierarchical capacity detection strategies to achieve dynamic modeling and management of the route network. Use the rule engine to monitor segment traffic in real time and execute predefined conflict avoidance rules; use optimization algorithms to dynamically adjust route allocation to maximize airspace utilization; use machine learning to predict congested nodes based on historical data and optimize long-term route planning. The route design steps include obtaining the target longitude and latitude, converting the route UTM longitude and latitude into two-dimensional coordinates, instantiating main roads, branch roads, and intersections, calculating the length of each type of route, and establishing inherent connections between routes.

[0123] Planning Layer: Considering multiple constraints, a depth-first search (DFS) combined with specific strategies is used to explore potential paths, obtaining the routes, nodes, and intersections required for flight planning, thereby providing efficient and effective routes. Depth-first search begins by selecting a starting point and then retracing when no further progress is possible until all connected vertices have been traversed. A stack is used to store access path information.

[0124] The decision-making layer predicts flight conflicts and dynamically adjusts aircraft takeoff times to ensure that there are no time conflicts at key intersections on the flight path and that minimum safety interval requirements are met. The specific steps include path planning, obtaining all possible flight paths, and terminating with an error if no valid path exists; preparing for conflict detection, extracting the intersection occupancy time periods of the existing route network and the key intersections of each path; calculating the timetable, calculating the theoretical time to reach each intersection based on aircraft speed and path length, and generating a time window; performing conflict detection and adjustment, checking whether the time window overlaps with existing occupied time periods. If there is a conflict, the takeoff time is postponed, with a maximum of max_attempts attempts. If a path is conflict-free, the adjusted time and path are immediately returned; setting termination conditions, returning the adjusted time if a conflict-free path is found, and throwing a TimeoutError if all paths conflict.

[0125] Scheduling layer: Genetic algorithm is used to solve the flight mission scheduling problem at multi-channel airports.

[0126] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will readily appreciate that other variations or modifications based on the above descriptions are possible. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A method for intelligent pre-flight scheduling of UAVs based on fixed routes, characterized in that: include: Step S1, constructing an airspace route topology structure based on route nodes and route segment attributes to form a route relationship network; Step S2, obtaining the routes, nodes and intersection points required for the flight plan based on the route relationship network and the flight plan submitted by the user; Step S3: predict flight conflicts and dynamically adjust takeoff time to ensure that there is no time conflict between the start and end points on the flight path and that the minimum safety interval requirement is met; Step S4: Schedule the flight missions so that all flight missions are completed within the requested time window and meet the minimum takeoff interval requirements.

2. The method according to claim 1, characterized in that In step S1, the steps for generating the route relationship network are: Get the UTM longitude and latitude of the route; Convert the route UTM longitude and latitude into two-dimensional coordinates; Obtain information on main roads, branch roads and intersections; Calculate the length of each route; Establish inter-route linkages.

3. The method according to claim 2, characterized in that In step S2, a depth-first search is used to obtain all possible flight paths, and the routes, nodes, and intersection points required for the flight plan are obtained from all possible flight paths.

4. The method according to claim 3, characterized in that Step S3 specifically includes: Get the time period occupied by the intersection point in the route relationship network; Get the starting and ending points of each path; Calculate the theoretical time to reach each intersection based on the flight speed and path length, and generate a time window for reaching each intersection based on the theoretical time; Check whether the time window of each intersection conflicts with the occupied time period; if there is a conflict, postpone the departure time; if there is no conflict, return the adjusted time and path.

5. The method according to claim 1, wherein In step S4, flight mission scheduling includes: A flight mission scheduling model is established, and the model objective function is: Fitness=Total Priority-α*Time Range-β*Path Diversity Among them, Fitness is the fitness, Total Prioritory is the sum of all priorities assigned to the UAV, TimeRange is the flight mission duration, Path Diversity is the difference in flight missions between routes, and α and β are weight coefficients.

6. The method according to claim 5, characterized in that Genetic algorithms are used to find the optimal solution for the flight mission scheduling model, specifically including: Individuals are represented as routes and schedules for all flight missions; The population is initialized to randomly assign routes and schedule takeoff times in sequence; Selection operations use tournament selection; The crossover operation is a two-point crossover, and the timing of the offspring is adjusted; The mutation operation is to randomly change the flight route allocation of the flight mission and adjust the time; Fitness evaluation uses the objective function.

7. The method according to claim 1, characterized in that In step S1, a segmented structure and hierarchical capacity detection strategy are used to establish a route relationship network.

8. The method according to claim 1, characterized in that In step S1, based on historical UAV flight trajectory data, the time period occupied by the intersection point is predicted through machine learning.

9. An intelligent pre-flight scheduling system for drones based on fixed routes, characterized in that: The system comprises: Road network layer, generating route relationship network based on route node and segment attributes; The planning layer obtains the routes, nodes, and intersections required for the flight plan based on the route relationship network and the flight plan submitted by the user; The decision-making layer predicts flight conflicts and dynamically adjusts takeoff times to ensure that there are no time conflicts between the start and end points of the flight path and that minimum safety interval requirements are met. The scheduling layer schedules flight missions so that all flight missions are completed within the requested time window and meet the minimum take-off interval requirements.

10. The system according to claim 9, characterized in that In the route network layer, the steps for generating the route relationship network are as follows: Get the UTM longitude and latitude of the route; Convert the route UTM longitude and latitude into two-dimensional coordinates; Obtain information on main roads, branch roads and intersections; Calculate the length of each route; Establishing inter-route linkages; In the planning layer, a depth-first search is used to obtain all possible flight paths, and the routes, nodes, and intersections required for the flight plan are obtained from all possible flight paths; At the decision-making level, predicting flight conflicts and dynamically adjusting takeoff times specifically include: Get the time period occupied by the intersection point in the route relationship network; Get the starting and ending points of each path; Calculate the theoretical time to reach each intersection based on the flight speed and path length, and generate a time window for reaching each intersection based on the theoretical time; Check whether the time window of each intersection point conflicts with the occupied time period; if there is a conflict, postpone the departure time; if there is no conflict, return the adjusted time and path; In the scheduling layer, flight mission scheduling includes: A flight mission scheduling model is established, and the model objective function is: Fitness=Total Priority-α*Time Range-β*Path Diversity Among them, Fitness is the fitness, Total Prioritory is the sum of all priorities assigned to the UAV, TimeRange is the flight mission duration, Path Diversity is the difference in flight missions between routes, and α and β are weight coefficients.

Images