Traffic light control method and device suitable for low-altitude multi-aircraft mixed take-off and landing field

Abstract

The invention relates to takeoff and landing field management and control, in particular to a traffic light management and control method and device suitable for a low-altitude multi-aircraft mixed takeoff and landing field, an airspace is divided into a plurality of three-dimensional grid units, the three-dimensional grid units suitable for flight are represented by green, and the three-dimensional grid units not suitable for flight are represented by red; three-dimensional grid units passed by the manned aircraft within a certain time period are prompted in a striking mode, it is determined that the corresponding airspace is a strict no-fly zone within a certain time period, and an improved fast and adaptive perception planning algorithm is adopted to plan a flight path for the unmanned aerial vehicle; according to the technical scheme, the defects that in the prior art, visual and refined expression is lacked in the aspect of airspace state management, and the requirements of a low-altitude multi-aircraft mixed take-off and landing field for flight safety and efficient operation are difficult to meet in the aspect of unmanned aerial vehicle flight path planning can be effectively overcome.

Description

Technical Field

[0001] The present invention relates to take-off and landing field control, and in particular to a traffic light control method and equipment applicable to a low-altitude multi-aircraft mixed take-off and landing field. Background Art

[0002] In recent years, aviation activity in the low-altitude sector has become increasingly active, with mixed takeoffs and landings of various aircraft, including drones and manned aircraft, becoming increasingly common. The low-altitude airspace environment is complex and ever-changing, with numerous factors impacting flight safety and efficiency.

[0003] Traditional methods for airspace status management lack intuitive and refined representation, making it difficult to meticulously demarcate airspace and clearly identify suitable and unsuitable airspace. Furthermore, there's no effective way to clearly define no-fly zones and time periods for manned aircraft. This makes it difficult for aircraft operators to accurately and timely understand airspace conditions, increasing the risk of flight conflicts and safety incidents.

[0004] When it comes to UAV flight path planning, the increasing number of low-altitude aircraft has led to large-scale, time-varying airspace, coupled with uncertainty in trajectory prediction for dynamic objects. Existing path planning algorithms have significant shortcomings in handling these complex situations: Firstly, they struggle to efficiently detect conflicts in large-scale, time-varying airspace and cannot generate airspace conflict pairs in a timely manner, resulting in potential conflict risks in the planned paths. Secondly, the uncertainty in trajectory prediction for dynamic objects is not fully considered during path planning, making the generated trajectories unfeasible and unsafe in the face of dynamic changes during actual flight, making it difficult to meet the flight safety and efficient operation requirements of low-altitude, multi-aircraft mixed takeoff and landing sites. Summary of the Invention

[0005] In response to the above-mentioned shortcomings of the existing technology, the present invention provides a traffic light control method and equipment suitable for low-altitude multi-aircraft mixed take-off and landing fields, which can effectively overcome the defects of the existing technology in lacking intuitive and refined expression in airspace status management, and difficulty in meeting the flight safety and efficient operation requirements of low-altitude multi-aircraft mixed take-off and landing fields in terms of UAV flight path planning.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A traffic light control method suitable for low-altitude, multi-aircraft mixed takeoff and landing sites divides the airspace into multiple three-dimensional grid cells, displays those suitable for flight in green, and those unsuitable in red. Three-dimensional grid cells that manned aircraft pass through within a certain time period are highlighted, clearly indicating that the corresponding airspace is a strict no-fly zone within a certain period of time. An improved fast and adaptive perception planning algorithm is used to plan the flight path for the drone. Among them, in improving the fast and adaptive perception planning algorithm: A space-time grid airspace conflict detection algorithm is used to handle conflict detection in large-scale time-varying airspaces. The vector method with high computational complexity is converted to set operations between grids with low computational complexity. The airspace conflict pair set is quickly generated. Based on the current position and target position of the UAV, as well as the airspace conflict pair set and the predicted trajectory of dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction, and the flight path is generated by combining the uncertainty-aware trajectory generation method.

[0007] Preferably, the method of planning a flight path for a UAV using an improved fast and adaptive perception planning algorithm includes: S1. The drone continuously collects real-time environmental data of the airspace surrounding the take-off and landing area and updates the status of the pre-divided three-dimensional grid cells; S2. Segment the collected point cloud data to identify static objects and dynamic objects, and track the trajectory of dynamic objects; S3. Use a space-time grid spatial conflict detection algorithm to handle conflict detection in large-scale time-varying airspaces. This algorithm converts the high-computational complexity of vector methods to low-computational complexity set operations between grids, quickly generates sets of airspace conflict pairs, and performs priority assessment on each set of airspace conflict pairs. S4. Based on the UAV's current position, target position, and the predicted trajectories of the airspace conflict pair set and dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, a penalty term is introduced in the trajectory generation process to deal with the uncertainty of the dynamic object trajectory prediction. The flight path is generated by combining the uncertainty-aware trajectory generation method. S5. Evaluate whether the current flight path has a collision risk based on the real-time predicted trajectory of the dynamic object. If there is a collision risk, perform adaptive replanning to generate a new flight path.

[0008] Preferably, the drone in S1 continuously collects real-time environmental data of the airspace surrounding the take-off and landing field, and updates the status of the pre-divided three-dimensional grid cells, including: S11. The drone continuously collects real-time environmental data, including point cloud data and image data, from the airspace surrounding the take-off and landing area through sensors such as lidar and cameras. S12. Based on the pre-divided three-dimensional grid cells, combined with real-time environmental data and airspace management information, update the status of each three-dimensional grid cell: The three-dimensional grid cells suitable for flight are represented in green, indicating that there are no obstacles in the corresponding airspace or the obstacles have no impact on flight; Three-dimensional grid cells that are not suitable for flight are shown in red, indicating that there are static obstacles in the corresponding airspace; The three-dimensional grid cells that manned aircraft pass through within a certain period of time are highlighted in red / flashing red to clearly indicate that the corresponding airspace is a strict no-fly zone within a certain period of time. Drones must avoid it to prioritize the flight safety of manned aircraft.

[0009] Preferably, in S2, the collected point cloud data is segmented to identify static objects and dynamic objects, and the trajectory of the dynamic objects is tracked, including: S21. Analyze the geometric features of the point cloud data, including distance, density, and curvature, and segment the point cloud data into different clusters to identify static and dynamic objects; S22. For the identified dynamic object, a data association method is used to associate the newly detected dynamic cluster with the existing tracker, and the trajectory of the dynamic object is tracked.

[0010] Preferably, in S22, for the identified dynamic object, a data association method is used to associate the newly detected dynamic cluster with the existing tracker, and the trajectory of the dynamic object is tracked, including: If the distance similarity and feature similarity between a dynamic cluster and a tracker meet the matching conditions, the dynamic cluster is assigned to the tracker, and the Kalman filter is used to predict the trajectory of the dynamic object.

[0011] Preferably, S3 uses a space-time grid spatial conflict detection algorithm to handle large-scale time-varying spatial domain conflict detection, converting from a vector method with high computational complexity to a set operation between grids with low computational complexity, quickly generating a set of spatial domain conflict pairs, and performing a priority evaluation on each set of spatial domain conflict pairs, including: S31. Perform multi-scale gridding on the airspace set surrounding the take-off and landing field to obtain the time code, altitude code, and plane dimension code of each airspace. Perform an intersection operation on any two airspaces in the order of time, altitude, and plane dimension, and generate an airspace conflict pair set based on the operation results. S32. Consider factors including the severity, time of occurrence, and type of dynamic objects involved in the conflict and determine the priority of each group of airspace conflict pairs.

[0012] Preferably, in S31, the airspace set around the take-off and landing field is divided into a multi-scale grid to obtain the time code, altitude code, and plane dimension code of each airspace, and an intersection operation is performed on any two airspaces in the order of time, altitude, and plane dimension. An airspace conflict pair set is generated based on the operation results, including: S311, input the airspace set around the take-off and landing field Airspaces={airspace1,airspace2,…,airspace N}, N is the number of airspaces in the airspace set Airspaces; S312, performing multi-scale gridding on the airspace set around the take-off and landing field to obtain the time code, altitude code, and plane dimension code of each airspace; S313, initialize the first round of airspace conflict pair set , traverse all airspace pairs in the airspace set Airspaces , the time encoding of space i Time encoding with spatial domain j For comparison, if , then add airspace i and airspace j to the first round of airspace conflict pair set ; S314: Initialize the second round of airspace conflict pair set , traverse the first round of airspace conflict pairs All airspace pairs in , N' is the set of airspace conflict pairs in the first round The number of airspaces in the space, encoding the height of the airspace m The height encoding of the airspace n For comparison, if , then add airspace m and airspace n to the second round of airspace conflict pair set ; S315. Initialize the final airspace conflict pair set , traverse the second round of airspace conflict pairs All airspace pairs in , N'' is the set of airspace conflict pairs in the second round The number of spaces in the space, encoding the plane dimension of the space x and the plane-dimensional encoding of the spatial domain y For comparison, if , then add airspace x and airspace y to the final airspace conflict pair set ; S316: Output the final airspace conflict pair set Conflict.

[0013] Preferably, in S4, a dynamically feasible and safe trajectory is directly generated based on the current position of the UAV, the target position, and the predicted trajectory of the airspace conflict pair set and the dynamic object. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of the trajectory prediction of the dynamic object. The flight path is generated by combining the uncertainty-aware trajectory generation method, including: Without the need for front-end path search and safe corridor generation, it directly generates dynamically feasible and safe trajectories based on the UAV's current position, target position, and predicted trajectories of airspace conflict pairs and dynamic objects; At the same time, during the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction. The flight path is generated by combining the uncertainty-aware trajectory generation method to increase the safe distance between the UAV and the dynamic object. Among them, the objective function F of trajectory generation is: ; In the above formula, f s is the item of expected smooth trajectory, f v +f a is the cost term with an upper limit on the velocity and acceleration of the desired trajectory, f c is the penalty term for the desired trajectory to be collision-free, 、 、 are weight coefficients, , is the trajectory point index of the UAV, J is the predicted trajectory point index of the dynamic object, is the distance between the trajectory point I of the UAV and the predicted trajectory point J of the dynamic object, is the weight coefficient.

[0014] Preferably, in S5, whether the current flight path has a collision risk is evaluated based on the real-time predicted trajectory of the dynamic object, and if there is a collision risk, adaptive replanning is performed to generate a new flight path, including: S51. Calculate the shortest distance between the current flight path and the dynamic object based on the real-time predicted trajectory of the dynamic object, and compare it with the safety threshold. If there is a collision risk, perform adaptive replanning. S52. When performing adaptive replanning, the algorithm calculates a transition position based on the real-time predicted trajectory of the dynamic object, pushes the UAV into free space, and regenerates a new flight path based on the UAV's target position, transition position, and the predicted trajectory of the airspace conflict pair set and dynamic objects.

[0015] A traffic light control device for a low-altitude, multi-aircraft mixed take-off and landing area, configured to execute the aforementioned traffic light control method for a low-altitude, multi-aircraft mixed take-off and landing area, includes a memory and a processor. The memory stores computer program instructions. When the computer program instructions are executed by the processor, the device is configured to perform the following operations: The airspace is divided into multiple three-dimensional grid cells, with those suitable for flight indicated in green and those unsuitable for flight indicated in red. Three-dimensional grid cells that manned aircraft pass through within a certain time period are highlighted, clearly indicating that the corresponding airspace is a strict no-fly zone within a certain period of time. An improved fast and adaptive perception planning algorithm is used to plan flight paths for drones. Among them, in improving the fast and adaptive perception planning algorithm: A space-time grid airspace conflict detection algorithm is used to handle conflict detection in large-scale time-varying airspaces. The vector method with high computational complexity is converted to set operations between grids with low computational complexity. The airspace conflict pair set is quickly generated. Based on the current position and target position of the UAV, as well as the airspace conflict pair set and the predicted trajectory of dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction, and the flight path is generated by combining the uncertainty-aware trajectory generation method.

[0016] Compared with the existing technology, the traffic light control method and equipment provided by the present invention, which is applicable to a low-altitude multi-aircraft mixed take-off and landing field, has the following beneficial effects: 1) Clear and intuitive airspace status: The airspace is divided into multiple three-dimensional grid cells, with suitable and unsuitable cells indicated in green and red, respectively. Furthermore, cells that manned aircraft pass through within a specific time period are highlighted as strictly no-fly zones. This intuitive airspace status display method enables aircraft operators to accurately and promptly understand airspace conditions, effectively reducing the risk of flight conflicts and safety accidents, and improving the safety of low-altitude flights. 2) Efficient Path Planning: An improved fast and adaptive perception planning algorithm is used to plan flight paths for drones. This algorithm utilizes a space-time grid-based airspace conflict detection algorithm to handle conflict detection in large, time-varying airspaces. This algorithm transforms the computationally complex vector method into a low-complexity set operation between grids, rapidly generating a set of airspace conflict pairs. Combining the drone's current position, target position, and predicted trajectories of dynamic objects, it directly generates a dynamically feasible and safe trajectory. This approach significantly improves the efficiency and accuracy of path planning, ensuring drone flight safety in complex low-altitude environments. 3) Effectively addressing uncertainty: During trajectory generation, a penalty term is introduced to address the uncertainty of dynamic object trajectory prediction. Combined with an uncertainty-aware trajectory generation method, the generated flight path is more robust, maintaining feasibility and safety in the face of dynamic changes in actual flight. This significantly enhances the adaptability and robustness of the UAV in complex low-altitude environments. 4) Improve operational efficiency: Through intuitive airspace status representation and efficient path planning, the present invention helps optimize the operational process of low-altitude multi-aircraft mixed take-off and landing fields, reducing waiting time and conflicts between aircraft, thereby improving the operational efficiency of the entire take-off and landing field. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0018] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 This is a flow chart of the process of planning the flight path for a UAV using the improved fast and adaptive perception planning algorithm in the present invention. DETAILED DESCRIPTION

[0019] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] Traffic light control method applicable to low-altitude multi-aircraft mixed take-off and landing fields, such as Figure 1 As shown, the airspace is divided into multiple three-dimensional grid cells. The three-dimensional grid cells suitable for flight are represented in green, and the three-dimensional grid cells unsuitable for flight are represented in red. The three-dimensional grid cells that manned aircraft pass through within a certain period of time are highlighted in a striking manner, clearly indicating that the corresponding airspace is a strict no-fly zone within a certain period of time. An improved fast and adaptive perception planning algorithm is used to plan the flight path for the UAV. Among them, in improving the fast and adaptive perception planning algorithm: A space-time grid airspace conflict detection algorithm is used to handle conflict detection in large-scale time-varying airspaces. The vector method with high computational complexity is converted to set operations between grids with low computational complexity. The airspace conflict pair set is quickly generated. Based on the current position and target position of the UAV, as well as the airspace conflict pair set and the predicted trajectory of dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction, and the flight path is generated by combining the uncertainty-aware trajectory generation method.

[0021] An improved fast and adaptive perception planning algorithm is used to plan the flight path for the drone, including: S1. The drone continuously collects real-time environmental data of the airspace surrounding the take-off and landing area and updates the status of the pre-divided three-dimensional grid cells; S2. Segment the collected point cloud data to identify static objects and dynamic objects, and track the trajectory of dynamic objects; S3. Use a space-time grid spatial conflict detection algorithm to handle conflict detection in large-scale time-varying airspaces. This algorithm converts the high-computational complexity of vector methods to low-computational complexity set operations between grids, quickly generates sets of airspace conflict pairs, and performs priority assessment on each set of airspace conflict pairs. S4. Based on the UAV's current position, target position, and the predicted trajectories of the airspace conflict pair set and dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, a penalty term is introduced in the trajectory generation process to deal with the uncertainty of the dynamic object trajectory prediction. The flight path is generated by combining the uncertainty-aware trajectory generation method. S5. Evaluate whether the current flight path has a collision risk based on the real-time predicted trajectory of the dynamic object. If there is a collision risk, perform adaptive replanning to generate a new flight path.

[0022] 1) In S1, the drone continuously collects real-time environmental data of the airspace surrounding the take-off and landing area and updates the status of the pre-divided three-dimensional grid cells, including: S11. The drone continuously collects real-time environmental data, including point cloud data and image data, from the airspace surrounding the take-off and landing area through sensors such as lidar and cameras. S12. Based on the pre-divided three-dimensional grid cells, combined with real-time environmental data and airspace management information, update the status of each three-dimensional grid cell: The three-dimensional grid cells suitable for flight are represented in green, indicating that there are no obstacles in the corresponding airspace or the obstacles have no impact on flight; Three-dimensional grid cells that are not suitable for flight are shown in red, indicating that there are static obstacles in the corresponding airspace; The three-dimensional grid cells that manned aircraft pass through within a certain period of time are highlighted in red / flashing red to clearly indicate that the corresponding airspace is a strict no-fly zone within a certain period of time. Drones must avoid it to prioritize the flight safety of manned aircraft.

[0023] 2) S2 performs point cloud segmentation on the collected point cloud data to identify static and dynamic objects and track the trajectory of dynamic objects, including: S21. Analyze the geometric features of the point cloud data, including distance, density, and curvature, and segment the point cloud data into different clusters to identify static and dynamic objects; S22. For the identified dynamic object, a data association method is used to associate the newly detected dynamic cluster with the existing tracker, and the trajectory of the dynamic object is tracked.

[0024] Specifically, in S22, for the identified dynamic object, a data association method is used to associate the newly detected dynamic cluster with the existing tracker, and the trajectory of the dynamic object is tracked, including: If the distance similarity and feature similarity between a dynamic cluster and a tracker meet the matching conditions, the dynamic cluster is assigned to the tracker, and the Kalman filter is used to predict the trajectory of the dynamic object.

[0025] 3) S3 uses a space-time grid spatial conflict detection algorithm to handle conflict detection in large-scale time-varying airspaces. It converts the high-computational complexity vector method to a low-computational complexity set operation between grids, quickly generates a set of airspace conflict pairs, and performs a priority assessment on each set of airspace conflict pairs, including: S31. Perform multi-scale gridding on the airspace set surrounding the take-off and landing field to obtain the time code, altitude code, and plane dimension code of each airspace. Perform an intersection operation on any two airspaces in the order of time, altitude, and plane dimension, and generate an airspace conflict pair set based on the operation results. S32. Consider factors including the severity, time of occurrence, and type of dynamic objects involved in the conflict and determine the priority of each group of airspace conflict pairs.

[0026] Specifically, in S31, the airspace set around the take-off and landing field is divided into multi-scale grids to obtain the time code, altitude code, and plane dimension code of each airspace. An intersection operation is performed on any two airspaces in the order of time, altitude, and plane dimension. Based on the operation results, an airspace conflict pair set is generated, including: S311, input the airspace set around the take-off and landing field Airspaces={airspace1,airspace2,…,airspace N}, N is the number of airspaces in the airspace set Airspaces; S312, performing multi-scale gridding on the airspace set around the take-off and landing field to obtain the time code, altitude code, and plane dimension code of each airspace; S313, initialize the first round of airspace conflict pair set , traverse all airspace pairs in the airspace set Airspaces , the time encoding of space i Time encoding with spatial domain j For comparison, if , then add airspace i and airspace j to the first round of airspace conflict pair set ; S314: Initialize the second round of airspace conflict pair set , traverse the first round of airspace conflict pairs All airspace pairs in , N' is the set of airspace conflict pairs in the first round The number of airspaces in the space, encoding the height of the airspace m Height encoding with airspace n For comparison, if , then add airspace m and airspace n to the second round of airspace conflict pair set ; S315. Initialize the final airspace conflict pair set , traverse the second round of airspace conflict pairs All airspace pairs in , N'' is the second round of airspace conflict set The number of spaces in the space, encoding the plane dimension of the space x and the plane-dimensional encoding of the spatial domain y For comparison, if , then add airspace x and airspace y to the final airspace conflict pair set ; S316: Output the final airspace conflict pair set Conflict.

[0027] 4) S4 directly generates a dynamically feasible and safe trajectory based on the UAV's current position, target position, and the predicted trajectories of the airspace conflict pair set and dynamic objects. During the trajectory generation process, a penalty term is introduced to address the uncertainty of dynamic object trajectory predictions. The flight path is generated by combining uncertainty-aware trajectory generation methods, including: Without the need for front-end path search and safe corridor generation, it directly generates dynamically feasible and safe trajectories based on the UAV's current position, target position, and predicted trajectories of airspace conflict pairs and dynamic objects; At the same time, during the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction. The flight path is generated by combining the uncertainty-aware trajectory generation method to increase the safe distance between the UAV and the dynamic object. Among them, the objective function F of trajectory generation is: ; In the above formula, f s is the item of expected smooth trajectory, f v +f a is the cost term with an upper limit on the velocity and acceleration of the desired trajectory, f c is the penalty term for the desired trajectory to be collision-free, 、 、 are weight coefficients, , is the trajectory point index of the UAV, J is the predicted trajectory point index of the dynamic object, is the distance between the trajectory point I of the UAV and the predicted trajectory point J of the dynamic object, is the weight coefficient.

[0028] 5) In S5, the current flight path is evaluated based on the real-time predicted trajectory of the dynamic object to determine whether there is a collision risk. If there is a collision risk, adaptive replanning is performed to generate a new flight path, including: S51. Calculate the shortest distance between the current flight path and the dynamic object based on the real-time predicted trajectory of the dynamic object, and compare it with the safety threshold. If there is a collision risk, perform adaptive replanning. S52. When performing adaptive replanning, the algorithm calculates a transition position based on the real-time predicted trajectory of the dynamic object, pushes the UAV into free space, and regenerates a new flight path based on the UAV's target position, transition position, and the predicted trajectory of the airspace conflict pair set and dynamic objects.

[0029] Based on the above-disclosed traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field, the technical solution of the present application further discloses a traffic light control device applicable to a low-altitude multi-aircraft mixed take-off and landing field, which is used to execute the above-disclosed traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field, including a memory and a processor, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the device is configured to perform the following operations: The airspace is divided into multiple three-dimensional grid cells, with those suitable for flight indicated in green and those unsuitable for flight indicated in red. Three-dimensional grid cells that manned aircraft pass through within a certain time period are highlighted, clearly indicating that the corresponding airspace is a strict no-fly zone within a certain period of time. An improved fast and adaptive perception planning algorithm is used to plan flight paths for drones. Among them, in improving the fast and adaptive perception planning algorithm: A space-time grid airspace conflict detection algorithm is used to handle conflict detection in large-scale time-varying airspaces. The vector method with high computational complexity is converted to set operations between grids with low computational complexity. The airspace conflict pair set is quickly generated. Based on the current position and target position of the UAV, as well as the airspace conflict pair set and the predicted trajectory of dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction, and the flight path is generated by combining the uncertainty-aware trajectory generation method.

[0030] In order to better illustrate the technical effect of using the space-time grid spatial domain conflict detection algorithm in the technical solution of this application to handle large-scale time-varying spatial domain conflict detection, the following three sets of comparative experiments are used to provide detailed descriptions: 1) Comparative Experiment 1: Small-Scale Scenario The space-time grid spatial conflict detection algorithm and the existing vector detection method are used to perform conflict detection on a small-scale time-varying airspace (including 30 airspaces). The experimental results are shown in Table 1: Table 1 Comparison of conflict detection results for 30 airspaces

[0031] As can be seen from Table 1, when there are 30 airspaces in the environment, the detection time of different methods is different. The vector detection method takes a longer time to detect. Both methods detect the same number of airspace conflict pairs.

[0032] 2) Comparative Experiment 2: Medium-Scale Scenario A. Using the space-time grid spatial conflict detection algorithm and the existing vector detection method to perform conflict detection on a medium-sized time-varying airspace (containing 100 airspaces), the experimental results are shown in Table 2: Table 2 Comparison of conflict detection results for 100 airspaces

[0033] B. Using the space-time grid spatial conflict detection algorithm and the existing vector detection method to perform conflict detection on a medium-sized time-varying airspace (containing 200 airspaces), the experimental results are shown in Table 3: Table 3 Comparison of conflict detection results for 200 airspaces

[0034] C. Using the space-time grid spatial conflict detection algorithm and the existing vector detection method to perform conflict detection on a medium-sized time-varying airspace (containing 400 airspaces), the experimental results are shown in Table 4: Table 4 Comparison of conflict detection results for 400 airspaces

[0035] It can be seen from Tables 2 to 4 that when the number of airspaces is 100, the detection time of the two methods is still below the second level, which can still meet the needs of rapid detection of airspace conflicts. However, the space-time grid airspace conflict detection algorithm has shown a more obvious advantage over the vector detection method. When the number of airspaces increases to 200, the time consumption of the vector detection method has reached the second level, while the time consumption of the space-time grid airspace conflict detection algorithm is still at the millisecond level. This is because the space-time grid airspace conflict detection algorithm can convert from the vector method with higher computational complexity to the set operation between grids with lower computational complexity, effectively solving the problem of increased computational complexity caused by the increase in the number of airspaces.

[0036] 3) Comparative Experiment 3: Large-Scale Scenario The space-time grid spatial conflict detection algorithm and the existing vector detection method are used to perform conflict detection on a large-scale time-varying airspace (including 1000 airspaces). The experimental results are shown in Table 5: Table 5 Comparison of conflict detection results for 1000 airspaces

[0037] As can be seen from Table 5, when dealing with conflict detection in large-scale time-varying airspace, the space-time grid airspace conflict detection algorithm has very obvious advantages over the vector detection method in terms of both the number of detected airspace conflict pairs and the detection time consumed.

[0038] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field, characterized by: The airspace is divided into multiple three-dimensional grid cells, with those suitable for flight indicated in green and those unsuitable for flight indicated in red. Three-dimensional grid cells that manned aircraft pass through within a certain time period are highlighted, clearly indicating that the corresponding airspace is a strict no-fly zone within a certain period of time. An improved fast and adaptive perception planning algorithm is used to plan flight paths for drones. Among them, in improving the fast and adaptive perception planning algorithm: A space-time grid airspace conflict detection algorithm is used to handle conflict detection in large-scale time-varying airspaces. The vector method with high computational complexity is converted to set operations between grids with low computational complexity. The airspace conflict pair set is quickly generated. Based on the current position and target position of the UAV, as well as the airspace conflict pair set and the predicted trajectory of dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction, and the flight path is generated by combining the uncertainty-aware trajectory generation method.

2. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 1 is characterized in that: The improved fast and adaptive perception planning algorithm is used to plan the flight path for the UAV, including: S1. The drone continuously collects real-time environmental data of the airspace surrounding the take-off and landing area and updates the status of the pre-divided three-dimensional grid cells; S2. Segment the collected point cloud data to identify static objects and dynamic objects, and track the trajectory of dynamic objects; S3. Use a space-time grid spatial conflict detection algorithm to handle conflict detection in large-scale time-varying airspaces. This algorithm converts the high-computational complexity of vector methods to low-computational complexity set operations between grids, quickly generates sets of airspace conflict pairs, and performs priority assessment on each set of airspace conflict pairs. S4. Based on the UAV's current position, target position, and the predicted trajectories of the airspace conflict pair set and dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, a penalty term is introduced in the trajectory generation process to deal with the uncertainty of the dynamic object trajectory prediction. The flight path is generated by combining the uncertainty-aware trajectory generation method. S5. Evaluate whether the current flight path has a collision risk based on the real-time predicted trajectory of the dynamic object. If there is a collision risk, perform adaptive replanning to generate a new flight path.

3. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 2 is characterized in that: In S1, the drone continuously collects real-time environmental data of the airspace surrounding the take-off and landing area and updates the status of the pre-divided three-dimensional grid cells, including: S11. The drone continuously collects real-time environmental data, including point cloud data and image data, from the airspace surrounding the take-off and landing area through sensors such as lidar and cameras. S12. Based on the pre-divided three-dimensional grid cells, combined with real-time environmental data and airspace management information, update the status of each three-dimensional grid cell: The three-dimensional grid cells suitable for flight are represented in green, indicating that there are no obstacles in the corresponding airspace or the obstacles have no impact on flight; Three-dimensional grid cells that are not suitable for flight are shown in red, indicating that there are static obstacles in the corresponding airspace; The three-dimensional grid cells that manned aircraft pass through within a certain period of time are highlighted in red / flashing red to clearly indicate that the corresponding airspace is a strict no-fly zone within a certain period of time. Drones must avoid it to prioritize the flight safety of manned aircraft.

4. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 2 is characterized in that: S2 performs point cloud segmentation on the collected point cloud data to identify static and dynamic objects and track the trajectory of dynamic objects, including: S21. Analyze the geometric features of the point cloud data, including distance, density, and curvature, and segment the point cloud data into different clusters to identify static and dynamic objects; S22. For the identified dynamic object, a data association method is used to associate the newly detected dynamic cluster with the existing tracker, and the trajectory of the dynamic object is tracked.

5. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 4 is characterized in that: In S22, for the identified dynamic objects, a data association method is used to associate the newly detected dynamic clusters with the existing trackers, and the trajectory of the dynamic objects is tracked, including: If the distance similarity and feature similarity between a dynamic cluster and a tracker meet the matching conditions, the dynamic cluster is assigned to the tracker, and the Kalman filter is used to predict the trajectory of the dynamic object.

6. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 2, characterized in that: S3 uses a space-time grid-based conflict detection algorithm to handle conflict detection in large-scale, time-varying airspaces. It converts the computationally complex vector method to a low-complexity set operation between grids, quickly generating sets of conflict pairs and performing priority assessments on each set of conflict pairs. This includes: S31. Perform multi-scale gridding on the airspace set surrounding the take-off and landing field to obtain the time code, altitude code, and plane dimension code of each airspace. Perform an intersection operation on any two airspaces in the order of time, altitude, and plane dimension, and generate an airspace conflict pair set based on the operation results. S32. Consider factors including the severity, time of occurrence, and type of dynamic objects involved in the conflict and determine the priority of each group of airspace conflict pairs.

7. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 6, characterized in that: In S31, the airspace set around the take-off and landing field is divided into multi-scale grids to obtain the time code, altitude code and plane dimension code of each airspace. Then, an intersection operation is performed on any two airspaces in the order of time, altitude and plane dimension. Based on the operation results, an airspace conflict pair set is generated, including: S311, input the airspace set around the take-off and landing field Airspaces={airspace1,airspace2,…,airspace N }, N is the number of airspaces in the airspace set Airspaces; S312, performing multi-scale gridding on the airspace set around the take-off and landing field to obtain the time code, altitude code, and plane dimension code of each airspace; S313, initialize the first round of airspace conflict pair set , traverse all airspace pairs in the airspace set Airspaces , the time encoding of space i Time encoding with spatial domain j For comparison, if , then add airspace i and airspace j to the first round of airspace conflict pair set ; S314: Initialize the second round of airspace conflict pair set , traverse the first round of airspace conflict pairs All airspace pairs in , N' is the set of airspace conflict pairs in the first round The number of airspaces in the space, encoding the height of the airspace m Height encoding with airspace n For comparison, if , then add airspace m and airspace n to the second round of airspace conflict pair set ; S315. Initialize the final airspace conflict pair set , traverse the second round of airspace conflict pairs All airspace pairs in , N'' is the second round of airspace conflict set The number of spaces in the space, encoding the plane dimension of the space x and the plane-dimensional encoding of the spatial domain y For comparison, if , then add airspace x and airspace y to the final airspace conflict pair set ; S316: Output the final airspace conflict pair set Conflict.

8. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 2, characterized in that: S4 directly generates a dynamically feasible and safe trajectory based on the drone's current position, target position, and the predicted trajectories of airspace conflict pairs and dynamic objects. During the trajectory generation process, a penalty term is introduced to address the uncertainty of dynamic object trajectory predictions. The flight path is generated by combining an uncertainty-aware trajectory generation method, including: Without the need for front-end path search and safe corridor generation, it directly generates dynamically feasible and safe trajectories based on the UAV's current position, target position, and predicted trajectories of airspace conflict pairs and dynamic objects; At the same time, during the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction. The flight path is generated by combining the uncertainty-aware trajectory generation method to increase the safe distance between the UAV and the dynamic object. Among them, the objective function F of trajectory generation is: ； In the above formula, f s is the item of expected smooth trajectory, f v +f a is the cost term with an upper limit on the velocity and acceleration of the desired trajectory, f c is the penalty term for the desired trajectory to be collision-free, 、 、 are weight coefficients, , is the trajectory point index of the UAV, J is the predicted trajectory point index of the dynamic object, is the distance between the trajectory point I of the UAV and the predicted trajectory point J of the dynamic object, is the weight coefficient.

9. The traffic light control method applicable to a low-altitude multi-aircraft mixed take-off and landing field according to claim 2, characterized in that: In S5, the current flight path is evaluated based on the real-time predicted trajectory of dynamic objects to determine whether there is a collision risk. If there is a collision risk, adaptive replanning is performed to generate a new flight path, including: S51. Calculate the shortest distance between the current flight path and the dynamic object based on the real-time predicted trajectory of the dynamic object, and compare it with the safety threshold. If there is a collision risk, perform adaptive replanning. S52. When performing adaptive replanning, the algorithm calculates a transition position based on the real-time predicted trajectory of the dynamic object, pushes the UAV into free space, and regenerates a new flight path based on the UAV's target position, transition position, and the predicted trajectory of the airspace conflict pair set and dynamic objects.

10. Traffic light control equipment applicable to a low-altitude mixed take-off and landing field for multiple aircraft, used to implement the traffic light control method applicable to a low-altitude mixed take-off and landing field for multiple aircraft according to claim 1, characterized in that: The device includes a memory and a processor, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the device is configured to perform the following operations: The airspace is divided into multiple three-dimensional grid cells, with those suitable for flight indicated in green and those unsuitable for flight indicated in red. Three-dimensional grid cells that manned aircraft pass through within a certain time period are highlighted, clearly indicating that the corresponding airspace is a strict no-fly zone within a certain period of time. An improved fast and adaptive perception planning algorithm is used to plan flight paths for drones. Among them, in improving the fast and adaptive perception planning algorithm: A space-time grid airspace conflict detection algorithm is used to handle conflict detection in large-scale time-varying airspaces. The vector method with high computational complexity is converted to set operations between grids with low computational complexity. The airspace conflict pair set is quickly generated. Based on the current position and target position of the UAV, as well as the airspace conflict pair set and the predicted trajectory of dynamic objects, a dynamically feasible and safe trajectory is directly generated. At the same time, in the trajectory generation process, a penalty term is introduced to deal with the uncertainty of dynamic object trajectory prediction, and the flight path is generated by combining the uncertainty-aware trajectory generation method.

Images