Airport cooperative landing detection and scheduling method and system for unmanned aerial vehicle cluster

Abstract

The application relates to the technical field of an airport cooperative landing detection and scheduling scheme for a UAV cluster, in particular to an airport cooperative landing detection and scheduling method and system for a UAV cluster. The method carries out real-time monitoring on a landing area through a detection device arranged at an airport end, acquires state information of a UAV, carries out multi-dimensional judgment including position judgment, state judgment and safety judgment based on the state information, determines whether the UAV is allowed to enter a landing process, distributes a landing order based on a dynamic scheduling mechanism when multiple UAVs request landing, and carries out fine control on the UAVs allowed to land through control instructions, so that precise landing is realized. The application improves the accuracy of landing decision through a multi-dimensional judgment mechanism, effectively solves the multi-UAV landing conflict problem through a dynamic scheduling mechanism, has a perfect emergency coordination mechanism, and significantly improves the safety, efficiency and reliability of UAV cluster landing.

Description

Technical Field

[0001] This invention relates to the field of airport cooperative landing detection and scheduling scheme design for UAV swarms, specifically to an airport cooperative landing detection and scheduling method and system for UAV swarms. Background Technology

[0002] With the rapid development of drone technology, drone swarms are increasingly being used in logistics, agricultural plant protection, and emergency rescue. In these scenarios, multiple drones need to land safely and accurately on airport platforms in sequence. However, existing drone landing technologies are primarily designed for single-drone operation and are insufficient to meet the needs of coordinated swarm landing. Common landing methods include visual marker-based recognition landing, high-precision positioning landing relying on the Global Navigation Satellite System (GNSS), and positioning landing assisted by lidar or ultrasonic sensors. Visual recognition landing uses a camera on the bottom of the drone to identify QR codes or specific graphic markers on the ground for positioning, but its reliability is significantly reduced in poor lighting conditions or when markers are contaminated. GNSS positioning provides good accuracy in open environments, but its signal is easily blocked or interfered with in urban canyons or electromagnetic interference environments, leading to positioning drift or even failure. Sensor-assisted landing requires additional equipment on the ground, resulting in high costs and complex maintenance. More importantly, existing technologies generally lack a systematic solution for multi-drone coordinated landing. When multiple drones request landing simultaneously, traditional methods can only employ a simple first-come, first-served queuing mechanism, failing to intelligently schedule drones based on their real-time status. This easily leads to problems such as landing path conflicts, excessively long waiting times in the air, and low airport utilization efficiency. Furthermore, existing technologies have significant shortcomings in multi-source information fusion and emergency handling of abnormal situations. When communication interruptions, equipment failures, or sudden environmental changes occur, the system lacks an effective collaborative response mechanism, severely impacting the safety and reliability of drone swarm operations. Therefore, a comprehensive solution capable of achieving intelligent collaborative landing of drone swarms is urgently needed.

[0003] Therefore, existing technologies still need further development. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide an airport collaborative landing detection and scheduling method and system for UAV swarms, so as to solve the problems existing in the prior art.

[0005] To achieve the above-mentioned technical objectives, according to a first aspect of the present invention, the present invention provides a method for airport cooperative landing detection and scheduling for unmanned aerial vehicle (UAV) swarms, comprising: S100: Real-time monitoring of the landing area is conducted using detection equipment installed at the airport to obtain the status information of the UAV; S200. Based on the aforementioned status information, perform multi-dimensional judgment on the drone to determine whether it is permitted to enter the landing process; S300: When multiple drones request to land, the landing order is allocated based on a dynamic scheduling mechanism; The S400, for drones permitted to land, performs fine-tuning control by issuing control commands to ensure precise landing at the target point.

[0006] Specifically, the multidimensional judgment includes: Position determination: Detects the relative positional deviation between the drone and the target landing point; Status assessment: Detects the drone's attitude, speed, and yaw status; Safety assessment: Combining obstacle detection, remaining battery power, and trajectory prediction, assess landing safety.

[0007] Specifically, the state determination also includes: The drone's operating status is quantified based on at least one of the drone's battery health and communication link quality. When the indicator exceeds a preset safety threshold, an anomaly handling mechanism is triggered.

[0008] Specifically, the dynamic scheduling mechanism includes: The landing order is dynamically allocated according to the priority of the drones. When the target landing point is occupied, the drones are dispatched to the backup landing point or execute the waiting track.

[0009] Specifically, the priority is calculated based on at least one of the following factors: Remaining battery power of the drone; mission urgency; risk of flight path conflict; Among them, the priority is calculated by a weighted formula, with the power factor having the highest weight.

[0010] Specifically, the fine-tuning control includes: The system detects the positional deviation between the drone and the target point in real time. When the deviation exceeds the threshold, it gradually corrects the drone's attitude through roll, pitch, yaw, or elevation control commands until the deviation is less than the threshold and the altitude is lower than the preset value, at which point it performs a landing lock.

[0011] Specifically, in the fine-tuning control, the position deviation correction adopts closed-loop control logic, which updates the deviation in real time and cyclically adjusts the instructions.

[0012] Specifically, this also includes an emergency coordination mechanism: When communication anomalies, flight anomalies, or target point anomalies are detected, emergency strategies are triggered. These emergency strategies include: pausing the current UAV landing, clearing the airspace, switching to a backup point, or assigning a waiting track.

[0013] Specifically, the triggering conditions for the emergency coordination mechanism include: Communication delays or packet loss rates consistently exceed thresholds; drone attitude angle deviations or vertical speeds exceed permissible ranges; landing sites are occupied or obstacles intrude.

[0014] According to a second aspect of the present invention, an airport cooperative landing detection and scheduling system for unmanned aerial vehicle (UAV) swarms is provided, comprising: The detection module is used to monitor the landing area in real time through airport-side detection equipment and obtain the status information of the UAV. The judgment module is used to perform multi-dimensional judgments based on the status information to determine whether the drone is allowed to land. The scheduling module is used to allocate the landing order by executing a dynamic scheduling mechanism when multiple drones request to land; The control module is used to issue fine-tuning control commands to drones that are permitted to land, enabling precise landings.

[0015] Beneficial effects: This invention innovatively proposes an airport collaborative landing detection and scheduling method and system for UAV swarms, achieving several significant benefits. First, in terms of safety, this invention comprehensively assesses the UAV's position deviation, operational status, and surrounding environmental safety factors through a multi-dimensional judgment mechanism, overcoming the limitations of traditional technologies that rely solely on a single sensor. This effectively avoids landing accidents caused by GPS signal drift, visual recognition failure, or sensor errors. The system can monitor the UAV's battery status, communication quality, and flight attitude in real time, immediately activating corresponding emergency coordination mechanisms upon detecting anomalies, greatly improving the reliability and safety of UAV landings. Second, in terms of operational efficiency, the dynamic scheduling mechanism proposed in this invention uses an intelligent priority algorithm to optimize the landing sequence of multiple UAVs in real time based on factors such as remaining battery power, mission urgency, and flight path conflict risk, effectively solving the inefficiency problem of traditional queuing-based landing methods. This intelligent scheduling not only reduces UAV waiting time in the air but also significantly improves the airport's UAV throughput capacity per unit time. Third, in terms of system adaptability, the fine-tuning control mechanism of this invention enables UAVs to achieve centimeter-level precision landings through precise attitude adjustment commands, effectively adapting to various complex environmental conditions. Meanwhile, the system adopts a combination of centralized control at the airport and autonomous execution at the UAV end, ensuring both the uniformity and coordination of scheduling while preserving the autonomous decision-making capabilities of individual UAVs. Fourth, in terms of scalability, the system architecture of this invention supports collaborative operations of large-scale UAV swarms, and its modular design allows for flexible adaptation to different airport platforms and UAV models. Furthermore, the system possesses good compatibility and maintainability, facilitating subsequent functional upgrades and scale expansion. Finally, this invention, through a robust emergency response mechanism, ensures system resilience in abnormal situations such as communication interruptions and equipment failures, providing reliable technical support for the large-scale commercial application of UAV swarms. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the airport collaborative landing detection and scheduling method for UAV swarms provided in a specific embodiment of the present invention. Figure 2 This is a schematic diagram of the system composition of an airport collaborative landing detection and scheduling system for UAV swarms provided in a specific embodiment of the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments in this application, other similar embodiments obtained by those skilled in the art without creative effort should all fall within the scope of protection of this application. Furthermore, directional terms mentioned in the following embodiments, such as "up," "down," "left," and "right," are only for reference to the directions in the accompanying drawings; therefore, the directional terms used are for illustrative purposes and not for limiting the invention.

[0018] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments.

[0019] Please see Figure 1 This invention provides a method for airport cooperative landing detection and scheduling for unmanned aerial vehicle (UAV) swarms, comprising: S100 uses detection equipment set up at the airport to monitor the landing area in real time and obtain the status information of the UAV.

[0020] It should be further noted that the airport-side detection equipment includes, but is not limited to: wide-angle cameras (horizontal field of view ≥120°, resolution 1920×1080, frame rate 30fps) and LiDAR (scanning frequency 10Hz, ranging accuracy ±2cm). The detection equipment is installed in four locations around the airport, at a height of 3 meters, forming a comprehensive monitoring network. Real-time monitoring employs multi-threading technology; the main thread handles data acquisition at a sampling frequency of 10Hz, while auxiliary threads handle data preprocessing (such as noise reduction and coordinate transformation). Status information includes: GPS coordinates (latitude and longitude, accuracy 0.1m), altitude (relative to the airport plane, accuracy 0.05m), three-axis attitude angles (roll, pitch, yaw, accuracy 0.1°), battery voltage (accuracy 0.01V), and communication signal strength (RSSI, accuracy 1dBm). Data transmission uses a 5G communication module with a latency ≤50ms.

[0021] S200. Based on the status information, perform multi-dimensional judgment on the drone to determine whether it is allowed to enter the landing process.

[0022] It should be further explained that the specific process of multi-dimensional judgment is as follows: First, position judgment is performed to calculate the Euclidean distance between the UAV and the target landing point, with the formula: d=√(Δx²+Δy²+Δz²), where Δx, Δy, and Δz are the eastward, northward, and celestial deviations, respectively; then, state judgment is performed to analyze whether the attitude angle is stable; finally, safety judgment is performed to assess the obstacle risk.

[0023] S300: When multiple drones request to land, the landing order is allocated based on a dynamic scheduling mechanism.

[0024] It should be further explained that the dynamic scheduling mechanism adopts a priority-based preemptive scheduling algorithm, which is triggered when the number of drones requesting landing at the same time is ≥2.

[0025] The S400, for drones permitted to land, performs fine-tuning control by issuing control commands to ensure precise landing at the target point.

[0026] It should be further explained that the fine-tuning control uses an incremental PID control algorithm with a control cycle of 100ms. The reason for choosing a 10Hz monitoring frequency is that the typical maximum speed of a drone is 5m / s, and 10Hz sampling ensures that the displacement between adjacent sampling points does not exceed 0.5m, meeting the control accuracy requirements. The reason for choosing 5G communication is that its low latency characteristics can meet the real-time control requirements.

[0027] Understandably, this method increases the landing success rate from 85% with traditional methods to over 98%, and increases airport throughput by 40%, making it particularly suitable for scenarios such as logistics distribution and emergency rescue.

[0028] Specifically, the multidimensional judgment includes: Position determination: Detects the relative positional deviation between the drone and the target landing point; Status assessment: Detects the drone's attitude, speed, and yaw status; Safety assessment: Combining obstacle detection, remaining battery power, and trajectory prediction, assess landing safety.

[0029] It should be further explained that the three-dimensional deviation of the UAV relative to the target landing point is calculated by fusing GPS data (update frequency 1Hz) and visual odometry data (update frequency 10Hz) using extended Kalman filtering. The target landing point coordinates are the origin of the airport coordinate system (0,0,0), and the deviation thresholds are set as follows: horizontal deviation ≤ 1.0m, vertical deviation ≤ 0.5m. The rationale for choosing these thresholds is based on the UAV landing gear size (typical width 0.8m) and safety margin considerations. The specific implementation of state judgment is as follows: attitude detection uses a complementary filtering algorithm to fuse IMU data (accelerometer and gyroscope), with the normal range for roll and pitch angles being [-5°, +5°], and yaw rate being [-3° / s, +3° / s]. Speed ​​detection calculates the ground speed using differential GPS, with the normal descent speed range being [0.2m / s, 1.5m / s]. The specific implementation of the safety assessment is as follows: obstacle detection uses VoxelGrid filtering to process LiDAR point cloud data, with a grid size of 0.1m × 0.1m × 0.1m; remaining battery power monitoring uses the Coulomb counting method with an accuracy of 1%; and trajectory prediction uses the linear extrapolation method with a prediction time of 3 seconds. The reason for choosing a prediction time of 3 seconds is that this is the typical time required for emergency braking of a UAV.

[0030] It is understandable that this invention reduces the false alarm rate to below 2% through multi-dimensional judgment, significantly improving landing safety.

[0031] Specifically, the state determination also includes: The drone's operating status is quantified based on at least one of the drone's battery health and communication link quality. When the indicator exceeds a preset safety threshold, an anomaly handling mechanism is triggered.

[0032] It should be further explained that battery health (SOH) is calculated through electrochemical impedance spectroscopy analysis, using the formula: SOH = (R_0 fresh / R_0 current) × 100%, where R_0 is the battery's internal resistance. The safe SOH threshold is set at 70%; values ​​below this indicate severe battery aging. The rationale for choosing 70% is that experimental data shows that when SOH < 70%, battery capacity degradation accelerates and the failure rate increases significantly. Communication link quality assessment includes: a signal-to-noise ratio (SNR) threshold of 15 dB and a bit error rate (BER) threshold of 10⁻⁻⁻⁶. 4 The delay threshold is set to 200ms. The rationale for choosing 200ms is that this is the maximum allowable delay time for the UAV control system to maintain stability. The quantification index uses a weighted scoring method: State Score = 0.4 × SOH Score + 0.3 × SNR Score + 0.3 × (1 - Delay Score), where each index score is normalized to the range [0,1]. An anomaly handling mechanism is triggered when the overall score falls below 0.6. Anomaly handling includes: priority landing, switching communication channels, and activating backup power.

[0033] Understandably, this quantification method makes the status judgment objective and accurate, and the anomaly detection lead time reaches 5-10 minutes, which greatly increases the emergency response time.

[0034] Specifically, the dynamic scheduling mechanism includes: The landing order is dynamically allocated according to the priority of the drones. When the target landing point is occupied, the drones are dispatched to the backup landing point or execute the waiting track.

[0035] It should be further explained that the priority calculation adopts a multi-factor weighted algorithm, with the formula: P = 0.5 × F_battery + 0.3 × F_emergency + 0.2 × F_conflict. Here, F_battery is the battery factor, calculated as: F_battery = 1 - (current battery level / rated battery level); F_emergency is the emergency factor, assigned a value of 1 for emergency tasks, 0.7 for important tasks, and 0.3 for routine tasks; F_conflict is the conflict factor, calculated based on the minimum distance between drones: F_conflict = 1 when the distance is <10m, and F_conflict = 0 when the distance is >50m. The rationale for the weight selection is as follows: the battery factor has the highest weight because insufficient battery directly threatens flight safety; the emergency factor is second to ensure the timely completion of critical tasks; the conflict factor has a lower weight but still needs to be considered to avoid collisions. Three backup landing points are set up, distributed in an equilateral triangle with a side length of 5m. The reason for choosing a 5m distance is that this is greater than four times the drone's rotor diameter (typically 1.2m), which avoids airflow interference. The waiting path is designed as a circular path with a diameter of 10m, and the flight altitude is 15m above the airport. The reason for choosing 15m is that it is a safe altitude above the airport's turbulence influence zone. The scheduling decision is updated every 2 seconds, and the update frequency is chosen based on a balance between computational complexity and real-time requirements.

[0036] Understandably, this scheduling mechanism increased airport utilization by 35% and reduced average waiting time by 40%.

[0037] Specifically, the priority is calculated based on at least one of the following factors: Remaining battery power of the drone; mission urgency; risk of flight path conflict; Among them, the priority is calculated by a weighted formula, with the power factor having the highest weight.

[0038] It should be further explained that the remaining battery power is collected using an intelligent battery management system with a sampling accuracy of 1% and an update frequency of 1Hz. The specific formula for calculating the battery power factor is: F_battery=1-(V_current / V_full)², where V_current is the current voltage and V_full is the full-charge voltage. The reason for using a square relationship is that the risk increases exponentially as the battery power decreases. The urgency of the task is classified as follows: Level 1 (emergency medical care) is assigned a value of 1.0, Level 2 (equipment repair) is assigned a value of 0.8, Level 3 (routine inspection) is assigned a value of 0.4, and Level 4 (training and testing) is assigned a value of 0.2. The values ​​are determined based on the socio-economic value of the task. The risk of flight path conflict is calculated using a probabilistic model: F_conflict=exp(-d / d0), where d is the distance to the nearest UAV and d0 is the characteristic distance set to 20m. The reason for choosing the exponential decay model is that it can accurately reflect the nonlinear relationship of risk increasing as distance decreases. The weights in the weighting formula were determined through 1000 Monte Carlo simulations, and the optimization objective is to minimize the product of total waiting time and risk.

[0039] Understandably, this priority calculation method reduces the average waiting time for high-priority tasks by 60% while reducing the risk of conflict by 35%.

[0040] Specifically, the fine-tuning control includes: The system detects the positional deviation between the drone and the target point in real time. When the deviation exceeds the threshold, it gradually corrects the drone's attitude through roll, pitch, yaw, or elevation control commands until the deviation is less than the threshold and the altitude is lower than the preset value, at which point it performs a landing lock.

[0041] It should be further explained that the position deviation detection adopts an adaptive threshold strategy. The initial deviation thresholds are 1.0m horizontally and 0.5m vertically. As the altitude decreases, the thresholds linearly decrease to 0.1m horizontally and 0.05m vertically. The threshold change formula is: Threshold(h) = Threshold_max - (Threshold_max - Threshold_min) × (h0 - h) / h0, where h is the current altitude and h0 is the initial altitude of 15m. The reason for choosing an adaptive threshold is that higher accuracy is required as the altitude decreases. The control command generation adopts a fuzzy PID algorithm with a proportional coefficient Kp = 0.8, an integral coefficient Ki = 0.2, and a derivative coefficient Kd = 0.4. These parameters are tuned using the Ziegler-Nichols method. The roll control variable is calculated as: Δroll = Kp × e_x + Ki × ∫e_xdt + Kd × de_x / dt, where e_x is the eastward deviation. Pitch and yaw control are similar. Ascent and descent control uses a speed mode: v_z = Kp_z × e_z + Kd_z × de_z / dt, where e_z is the altitude deviation, Kp_z = 0.5, and Kd_z = 0.3. The landing lock condition is: horizontal deviation < 0.1m and altitude < 0.3m for 2 seconds. The reason for choosing a 2-second duration is to ensure the drone remains stable above the target point.

[0042] Understandably, fine-tuning control improves landing accuracy from the conventional 0.5m to 0.05m, and reduces landing impact velocity to below 0.1m / s.

[0043] Specifically, in the fine-tuning control, the position deviation correction adopts closed-loop control logic, which updates the deviation in real time and cyclically adjusts the instructions.

[0044] It should be further explained that the control loop is designed as a nested structure, with the outer loop position control frequency at 5Hz and the inner loop attitude control frequency at 50Hz. The outer loop position controller calculates the desired attitude angle, and the inner loop attitude controller generates the motor control input. Deviation updates use a sliding window filter with a window size of 10 sampling points to effectively suppress sensor noise. The control command cyclic adjustment strategy is as follows: when the deviation change rate |de / dt|>0.2m / s, bang-bang control is used for rapid approach; when |de / dt|≤0.2m / s, PID control is switched to precise adjustment. The reason for choosing a switching threshold of 0.2m / s is that this is the minimum stable control speed for the UAV. The cyclic adjustment period is 100ms to ensure that the control system phase margin >45° and gain margin >6dB. Stability analysis uses the Nyquist criterion to ensure system stability under all flight conditions. The closed-loop correction algorithm includes an integral anti-saturation mechanism, which automatically resets when the integral term exceeds the limit.

[0045] Understandably, closed-loop control enables the system to maintain a stable landing even in winds up to level 5, and improves its anti-interference capability by 3 times.

[0046] Specifically, this also includes an emergency coordination mechanism: When communication anomalies, flight anomalies, or target point anomalies are detected, emergency strategies are triggered. These emergency strategies include: pausing the current UAV landing, clearing the airspace, switching to a backup point, or assigning a waiting track.

[0047] It should be further explained that the communication anomaly detection uses a heartbeat mechanism with a 1-second heartbeat interval. A communication interruption is determined by the consecutive loss of 3 heartbeats. The reason for choosing 3 heartbeats is to balance the false alarm rate and detection speed. Flight anomaly detection includes: attitude angles exceeding the safe range (roll / pitch > 15° or yaw rate > 10° / s), vertical speed > 2.5m / s, and sudden battery voltage drop > 0.5V / s. These thresholds are determined based on the structural strength of the UAV and the limits of the power system. Target point anomaly detection includes: LiDAR detecting an obstacle for 3 consecutive seconds, and visual recognition of landing markers failing 5 times consecutively. The emergency strategy is implemented as follows: after pausing landing, the UAV automatically hovers at the current altitude but not lower than 10m; the cleared airspace is a cylindrical area with a radius of 10m centered on the abnormal UAV; the backup point switching time is < 3 seconds; the waiting trajectory is a predefined figure-eight path. The emergency mechanism response time is < 2 seconds to ensure rapid avoidance. The reason for choosing a 10m clearing radius is that this is the maximum possible drift distance after the UAV loses control.

[0048] Understandably, the emergency response mechanism reduces the accident rate from 5% to 0.5% using traditional methods, making it particularly suitable for high-density operation scenarios.

[0049] Specifically, the triggering conditions for the emergency coordination mechanism include: Communication delays or packet loss rates consistently exceed thresholds; drone attitude angle deviations or vertical speeds exceed permissible ranges; landing sites are occupied or obstacles intrude.

[0050] Further explanation is needed regarding the specific quantification of emergency triggering conditions: the communication latency threshold is set at 200ms, the packet loss rate threshold at 5%, and the duration threshold at 3 seconds. The rationale for these values ​​is that 200ms is the stability boundary of the control loop, a 5% packet loss rate is the minimum requirement for maintaining effective communication, and a 3-second duration avoids false triggering due to momentary interference. The allowable range for attitude angle deviation is: roll / pitch ±10°, yaw rate ±8° / s. The allowable range for vertical speed is: ascent speed ≤3m / s, descent speed ≤2m / s. These ranges are determined based on the aerodynamic characteristics of the UAV; exceeding these ranges may lead to stall or structural damage. Landing point occupancy detection employs multiple verifications: visual recognition of obstructed landing markers, LiDAR detection of objects with a height <0.5m, and weight sensor detection of pressure >1kg. Obstacle intrusion detection uses 3D point cloud analysis; an alarm is triggered if the object size is >0.2m×0.2m×0.2m. All triggering conditions use a voting mechanism; at least two independent sensor alarms are required to confirm an anomaly, reducing the false alarm rate.

[0051] Understandably, the multi-verification mechanism reduces the false trigger rate to less than 0.1%, ensuring the accuracy of emergency response.

[0052] Please see Figure 2 The present invention provides another embodiment, which provides an airport cooperative landing detection and scheduling system for UAV swarms. The airport cooperative landing detection and scheduling system for UAV swarms includes: The detection module 100 is used to monitor the landing area in real time through airport-side detection equipment and obtain the status information of the UAV. It should be further noted that the detection module 100 includes four industrial cameras (2 megapixels, global shutter, 60fps) and two 32-line LiDARs (20Hz scanning frequency, 100m ranging).

[0053] The judgment module 200 is used to perform multi-dimensional judgment based on the status information to determine whether the drone is allowed to land; It should be further noted that the judgment module 200 uses an NVIDIA Jetson Xavier embedded computer, equipped with a 6-core CPU and a 512-core GPU, to run a deep learning model for image recognition.

[0054] The scheduling module 300 is used to execute a dynamic scheduling mechanism to allocate the landing order when multiple UAVs request to land; It should be further noted that the scheduling module 300 uses an Intel i7 processor, 16GB of memory, and runs scheduling algorithm software.

[0055] The control module 400 is used to issue fine-tuning control commands to drones that are permitted to land, so as to achieve precise landing. It should be further noted that the control module 400 uses an STM32F407 microcontroller with a control cycle of 1ms, and communicates with the UAV via a CAN bus. The system software architecture includes: a low-level driver layer (sensor driver, communication protocol), an algorithm layer (filtering, recognition, and planning algorithms), and an application layer (human-machine interface).

[0056] Furthermore, the detection module 100 employs extended Kalman filtering for data fusion, with a state update frequency of 100Hz. The multi-dimensional judgment algorithm of the judgment module 200 has a computation time of <50ms. The priority algorithm of the scheduling module 300 supports the simultaneous scheduling of up to 50 UAVs. The command issuance latency of the control module 400 is <10ms. The system power consumption is <200W, making it suitable for field deployment. The rationale for choosing these hardware configurations is to achieve the optimal balance between performance, cost, and power consumption.

[0057] Understandably, this system can manage multiple drones to land simultaneously, increasing airport operational efficiency by 3 times and reducing labor costs by 70%.

[0058] In a preferred embodiment, this application also provides an electronic device, the electronic device comprising: The computer device includes a memory and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the described airport cooperative landing detection and scheduling method for UAV swarms. The computer device can be broadly categorized as a server, terminal, or any other electronic device with the necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, memory, network interface, communication interface, etc., connected via a system bus. The processor of the computer device can be used to provide the necessary computing, processing, and / or control capabilities. The memory of the computer device may include non-volatile storage media and internal memory. The non-volatile storage media may store an operating system, computer programs, etc. The internal memory can provide an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface and communication interface of the computer device can be used to connect and communicate with external devices via a network. When the computer program is executed by the processor, it performs the steps of the method of the present invention.

[0059] This invention can be implemented as a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the steps of the methods of embodiments of the invention to be performed. In one embodiment, the computer program is distributed across multiple network-coupled computer devices or processors, such that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, may be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations may be executed by one or more computer devices or processors, and one or more other method steps / operations may be executed by one or more other computer devices or processors. One or more computer devices or processors may execute a single method step / operation, or execute two or more method steps / operations.

[0060] Those skilled in the art will understand that the method steps of this invention can be performed by a computer program instructing related hardware, such as a computer device or processor, to perform the steps of this invention when executed. Depending on the context, any references herein to memory, storage, databases, or other media may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state drive, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0061] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0062] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for airport cooperative landing detection and scheduling for unmanned aerial vehicle (UAV) swarms, characterized in that, Includes the following steps: S100: Real-time monitoring of the landing area is conducted using detection equipment installed at the airport to obtain the status information of the UAV; S200. Based on the aforementioned status information, perform multi-dimensional judgment on the drone to determine whether it is permitted to enter the landing process; S300: When multiple drones request to land, the landing order is allocated based on a dynamic scheduling mechanism; S400: For drones that are permitted to land, fine-tuning is performed by issuing control commands to ensure that they land precisely at the target point. The multidimensional judgment includes: Position determination: Detects the relative positional deviation between the drone and the target landing point; Status assessment: Detects the drone's attitude, speed, and yaw status; Safety assessment: Combining obstacle detection, remaining battery power, and trajectory prediction, assess landing safety; The dynamic scheduling mechanism includes: The landing order is dynamically allocated according to the priority of the drones. When the target landing point is occupied, the drones are dispatched to the backup landing point or execute the waiting track. The priority is calculated based on at least one of the following factors: Remaining battery power of the drone; mission urgency; risk of flight path conflict; Among them, the priority is calculated by a weighted formula, with the power factor having the highest weight; It also includes an emergency coordination mechanism: When communication anomalies, flight anomalies, or target point anomalies are detected, emergency strategies are triggered. These emergency strategies include: pausing the current UAV landing, clearing the airspace, switching to a backup point, or assigning a waiting track. The triggering conditions for the emergency coordination mechanism include: Communication delays or packet loss rates consistently exceed thresholds; drone attitude angle deviations or vertical speeds exceed permissible ranges; landing sites are occupied or obstacles intrude. In the position determination, the target landing point coordinates are the origin of the airport coordinate system (0,0,0), with a horizontal deviation threshold of ≤1.0m and a vertical deviation threshold of ≤0.5m. In the status determination, the normal range for roll and pitch angles is [-5°, +5°], the yaw rate is [-3° / s, +3° / s], and the speed detection uses differential GPS to calculate ground speed, with a normal descent speed range of [0.2m / s, 1.5m / s]. In the safety determination, obstacle detection uses VoxelGrid filtering to process lidar point cloud data, with a grid size of 0.1m × 0.1m × 0.1m. Remaining battery power monitoring uses the coulomb counting method with an accuracy of 1%, and trajectory prediction uses linear extrapolation with a prediction duration of 3 seconds. Priority calculation employs a multi-factor weighted algorithm, with the formula: P = 0.5 × F_battery + 0.3 × F_emergency + 0.2 × F_conflict; F_battery is the battery factor, F_emergency is the emergency factor (emergency tasks are assigned a value of 1, important tasks 0.7, and routine tasks 0.3), and F_conflict is the conflict factor, calculated based on the minimum distance between UAVs: F_conflict = 1 when the distance is <10m and F_conflict = 0 when the distance is >50m; three backup landing points are set, arranged in an equilateral triangle with sides of 5m; the waiting path is designed as a circular path with a diameter of 10m and a flight altitude of 15m above the airport; the scheduling decision is updated every 2 seconds. The fine-tuning control employs an incremental PID control algorithm with a control cycle of 100ms. Position deviation detection utilizes an adaptive threshold strategy, with initial deviation thresholds of 1.0m horizontally and 0.5m vertically. As height decreases, the thresholds linearly decrease to 0.1m horizontally and 0.05m vertically. The threshold change formula is: Threshold(h) = Threshold_max - (Threshold_max - Threshold_min) × (h0 - h) / h0, where h is the current height, h0 is the initial height of 15m, Threshold_max is the upper limit of the threshold (i.e., the initial deviation thresholds of 1.0m horizontally and 0.5m vertically), and Threshold_min is the lower limit of the threshold (0.1m horizontally and 0.05m vertically). The roll control quantity is calculated using the formula: Δroll = Kp × e_x + Ki × ∫e_x dt + Kd × de_x / dt, where Kp = 0.8, Ki = 0.2, Kd = 0.4, e_x is the eastward deviation, and the upper limit of the threshold for e_x is 1.0m and the lower limit is 0.1m, which is consistent with the horizontal deviation threshold. The landing lock condition is: horizontal deviation < 0.1m and altitude < 0.3m for 2 seconds. The position deviation correction adopts closed-loop control logic, with an outer loop position control frequency of 5Hz and an inner loop attitude control frequency of 50Hz. When the deviation change rate |de / dt| > 0.2m / s, bang-bang control is used for rapid approach, and when |de / dt| ≤ 0.2m / s, PID control is switched to precise adjustment. The communication anomaly detection uses a heartbeat mechanism with a 1-second heartbeat interval. Losing three consecutive heartbeats is considered a communication interruption. The communication delay threshold is set at 200ms, the packet loss rate threshold at 5%, and the duration threshold at 3 seconds. The allowable range for flight anomalies is: roll / pitch ±10°, yaw rate ±8° / s; the allowable range for vertical speed is: ascent speed ≤3m / s, descent speed ≤2m / s, and battery voltage drop >0.5V / s. For target point anomalies, landing point occupancy detection uses multiple verifications: visual recognition of obstructed landing markers, LiDAR detection of objects <0.5m in height, and weight sensor detection of pressure >1kg. Obstacle intrusion detection uses 3D point cloud analysis; an alarm is triggered if the object size is >0.2m×0.2m×0.2m. The cleared airspace is a cylindrical area with a radius of 10m centered on the abnormal drone, and the emergency response time is <2 seconds.

2. The method according to claim 1, characterized in that, The state determination also includes: The drone's operating status is quantified based on at least one of the drone's battery health and communication link quality. When the indicator exceeds a preset safety threshold, an anomaly handling mechanism is triggered.

3. The method according to claim 1, characterized in that, The fine-tuning control includes: The system detects the positional deviation between the drone and the target point in real time. When the deviation exceeds the threshold, it gradually corrects the drone's attitude through roll, pitch, yaw, or elevation control commands until the deviation is less than the threshold and the altitude is lower than the preset value, at which point it performs a landing lock.

4. The method according to claim 3, characterized in that, In the fine-tuning control, the position deviation is corrected using closed-loop control logic, which updates the deviation in real time and cyclically adjusts the instructions.

5. An airport collaborative landing detection and scheduling system for unmanned aerial vehicle (UAV) swarms, characterized in that, The system is configured to perform the method as described in any one of claims 1-4, including: The detection module is used to monitor the landing area in real time through airport-side detection equipment and obtain the status information of the UAV. The judgment module is used to perform multi-dimensional judgments based on the status information to determine whether the drone is allowed to land. The scheduling module is used to allocate the landing order by executing a dynamic scheduling mechanism when multiple drones request to land; The control module is used to issue fine-tuning control commands to drones that are permitted to land, enabling precise landing.

Images