Unmanned aerial vehicle intelligent landing assisting guiding method and system based on AI

Abstract

The invention discloses an AI-based unmanned aerial vehicle intelligent landing assisting and guiding method and system, and particularly relates to the technical field of low-altitude economic unmanned aerial vehicle control, and the AI-based unmanned aerial vehicle intelligent landing assisting and guiding method comprises the following steps: S1, carrying out the long-distance detection of medium and large unmanned aerial vehicles in a range of 5 kilometers through a phased array radar, therefore, the initial position, speed and course angle data of the unmanned aerial vehicle are obtained, and basic data are provided for subsequent landing guidance. S2, scanning an airborne laser enhancement ball of the unmanned aerial vehicle through laser target identification; and S3, driving the laser range finder by means of the high-precision two-dimensional mechanical tracking platform. According to the invention, the problems of insufficient long-distance perception capability, lack of legality verification, poor dynamic environment adaptability, weak multi-aircraft compatibility and the like in the existing unmanned aerial vehicle landing assisting technology can be effectively solved, and safe and accurate landing of medium and large unmanned aerial vehicles on temporary take-off and landing points or mobile platforms is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of low-altitude economy UAV control technology. More specifically, the present invention relates to an AI-based intelligent landing guidance method and system for UAVs. Background Art

[0002] At present, with the rapid development of the low-altitude economy, UAVs are increasingly widely used in many fields. However, the existing UAV landing technologies have many defects and are difficult to meet the actual needs: Insufficient long-distance sensing ability: Traditional radars have low detection accuracy for small and medium-sized UAVs more than 5 kilometers away. For large and medium-sized UAVs, accurate path planning cannot be carried out in advance, which severely limits the application of UAVs in some scenarios that require long-distance flight and precise landing; Lack of legality verification: Currently, there is no effective mechanism to quickly identify the identity and type of UAVs, which poses potential safety risks in some important areas or specific tasks. For example, the intrusion of illegal UAVs may interfere with normal operations or cause safety accidents; Poor adaptability to dynamic environments: The terrain of temporary takeoff and landing points is often complex and changeable, and at the same time, meteorological conditions are also unstable. Factors such as wind speed, precipitation, and visibility will all affect the landing of UAVs. The traditional flight control system mainly relies on preset programs and is difficult to make timely and effective adjustments in the face of these sudden interferences, easily leading to landing failures; Weak compatibility with multiple UAV models: There are significant differences in the landing principles and strategies between rotorcraft and fixed-wing aircraft. Existing systems are difficult to uniformly optimize and control different types of UAVs and cannot meet the requirements of diverse application scenarios.

[0003] To solve the above problems, a technical solution is provided now. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an AI-based intelligent landing guidance method and system for UAVs to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions: An AI-based intelligent landing guidance method for UAVs includes the following steps: Step S1, using a phased array radar to remotely detect large and medium-sized UAVs within a range of 5 kilometers, so as to obtain the initial position, speed, and heading angle data of the UAVs, providing basic data for subsequent landing guidance; Step S2, scanning the laser enhancement sphere carried by the UAV through laser target recognition; Step S3: Drive the laser rangefinder with the high-precision two-dimensional mechanical tracking platform to track the three-dimensional coordinates of the UAV in real time; Step S4: Integrate the terrain data of the temporary takeoff and landing point, real-time meteorological information, and the type of the UAV, deploy a multi-objective optimization algorithm to generate a dynamic landing trajectory and control instructions suitable for the UAV; Step S5: Transmit the generated trajectory instructions to the UAV through an encrypted data link to perform autonomous landing operations.

[0006] In a preferred embodiment, in Step S1: The phased array radar operates in the frequency band of 12 - 18 GHz, with a positioning accuracy of ≤ 5 m at a detection distance of 5 km, and supports simultaneous tracking of ≥ 20 targets, capable of meeting the monitoring requirements for multiple UAVs to land simultaneously.

[0007] In a preferred embodiment, in Step S2: The laser-enhanced sphere contains a unique code and type identification. Using the dual authentication technology of nano-level reflective coating and radio frequency coding, the reflectivity is ≥ 90% in harsh environments such as rain and fog, and the illegal UAV identification rate is ≥ 99.9%. It can accurately verify the legality of the UAV and classify it.

[0008] In a preferred embodiment, in Step S3: The accuracy is ≤ 0.1 m, and its motion state is obtained simultaneously. The two-dimensional mechanical tracking platform uses a high-torque servo motor and closed-loop feedback control, with a horizontal / pitch angle tracking accuracy of ≤ 0.01°, and a response delay of ≤ 10 ms, capable of accurately tracking the dynamic changes of the UAV.

[0009] In a preferred embodiment, in Step S4: For different types of UAVs, different landing strategies are adopted. The rotorcraft gives priority to ensuring the vertical landing accuracy, with an error of ≤ 0.2 m. The fixed-wing aircraft optimizes the glide angle and touchdown speed, with an error of ≤ 5%. The survey aircraft supports buffered landing when carrying heavy equipment, with an impact force of ≤ 5G.

[0010] In a preferred embodiment, in Step S4: The real-time meteorological information includes wind speed, precipitation, and visibility.

[0011] In a preferred embodiment, in Step S5: The encrypted data link supports dual-band redundant communication (L band + C band), with a communication interruption rate ≤ 0.1% in an electromagnetic interference environment. The flight control system of the UAV integrates on-board vision (such as infrared cameras, lidar) and ground commands, and generates the final control signal through the Extended Kalman Filter (EKF) to perform autonomous landing operations. At the same time, for gust disturbance situations, a feedforward-feedback composite control strategy is adopted to improve the attitude stability by more than 30%.

[0012] An AI-based intelligent landing guidance system for UAVs, used to implement the AI-based intelligent landing guidance method for UAVs described in any one of the above items, including a long-distance sensing module, a laser identification and ranging module, an AI management and control center module, and a UAV-side fusion controller module; The long-distance sensing module integrates a Ku-band phased array radar with a detection range of 5 - 10 kilometers, and a wide-angle optoelectronic tracking device, which is used to detect the position and status of the UAV at a long distance; The laser identification and ranging module includes a 532nm laser emitter, a high-sensitivity photodetector, and a mechanical tracking platform to achieve the legal verification and high-precision ranging of the UAV; The AI management and control center module is used to deploy edge computing nodes, run a trajectory generation model based on deep reinforcement learning, integrate multi-source data, and generate a dynamic landing trajectory and control instructions; The UAV-side fusion controller module integrates a visual SLAM module, an instruction parsing unit, and an anti-disturbance flight control algorithm to achieve the fusion of ground instructions and on-board information and control the landing process of the UAV.

[0013] The technical effects and advantages of the AI-based intelligent landing guidance method and system of the present invention: 1. It can detect UAVs at a distance of more than 5 kilometers with a detection rate ≥ 99%, and the illegal target interception rate reaches 100%, ensuring the safety of the landing area. The vertical error of the rotorcraft is ≤ 0.15m, and the landing speed error of the fixed-wing aircraft is ≤ 3%, meeting the high-precision landing requirements of different types of UAVs; 2. The time from target recognition to trajectory generation is ≤ 200ms, which can quickly adapt to a dynamic environment with a wind speed change ≥ 10m / s², improving the emergency handling ability of the system. Through multi-modal perception and dynamic path planning, it can adapt to the complex terrain and changing meteorological conditions of temporary takeoff and landing points, as well as the dynamic changes of mobile platforms. The same system can be compatible with the differential control requirements of different types of UAVs such as rotorcraft hovering landing and fixed-wing aircraft taxiing landing, and has wide applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flow schematic diagram of an AI-based intelligent landing guidance method of the present invention; Figure 2 This is a schematic structural diagram of an AI-based intelligent landing guidance system for drones according to the present invention. Specific Embodiments

[0015] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0016] Embodiment 1 The present invention proposes an AI-based intelligent landing guidance method and system for drones.

[0017] Figure 1 An AI-based intelligent landing guidance method for drones according to the present invention is given, including the following steps: Step S1, using a phased array radar to remotely detect medium and large drones within a range of 5 kilometers, thereby obtaining the initial position, speed, and heading angle data of the drones, providing basic data for subsequent landing guidance; Step S2, scanning the laser enhancement sphere carried by the drone through laser target recognition; Step S3, driving a laser rangefinder with a high-precision two-dimensional mechanical tracking platform to continuously track the three-dimensional coordinates of the drone; Step S4, integrating the terrain data of the temporary takeoff and landing point, real-time meteorological information, and the type of the drone, deploying a multi-objective optimization algorithm to generate a dynamic landing trajectory and control instructions suitable for the drone; Step S5, transmitting the generated trajectory instructions to the drone through an encrypted data link to perform an autonomous landing operation.

[0018] In step S1: The working frequency band of the phased array radar is 12 - 18 GHz. When the detection range is 5 kilometers, the positioning accuracy is ≤ 5 m, and it supports simultaneously tracking ≥ 20 targets, capable of meeting the monitoring requirements for multiple drones to land simultaneously.

[0019] In step S2: The laser enhancement sphere contains a unique code and type identifier. Using the dual authentication technology of nano-level reflective coating and radio frequency coding, the reflectivity is ≥ 90% in harsh environments such as rain and fog, and the illegal drone recognition rate is ≥ 99.9%. It accurately verifies the legality of the drone and classifies it.

[0020] It should be added that by combining optical reflection and radio frequency coding, millisecond-level verification of the drone's identity and model is achieved.

[0021] In step S3: The accuracy is ≤ 0.1 m. At the same time, its motion state is obtained. The two-dimensional mechanical tracking platform uses a high-torque servo motor and closed-loop feedback control. The horizontal / pitch angle tracking accuracy is ≤ 0.01°, and the response delay is ≤ 10 ms, which can accurately track the dynamic changes of the UAV.

[0022] It should be added that the problem of tracking jitter of small targets at a long distance is solved through a high-dynamic two-dimensional platform and laser ranging.

[0023] The two-dimensional mechanical platform uses a harmonic reducer + absolute encoder, with an angular resolution of 0.001° and a maximum rotation speed of 30° / s; The laser ranging pulse repetition frequency is 1 kHz, the ranging accuracy is ±2 cm @1 km, and it supports continuous tracking of moving targets.

[0024] In step S4: For different types of UAVs, different landing strategies are adopted. For rotorcraft, the vertical landing accuracy is given priority, with an error ≤ 0.2 m. For fixed-wing aircraft, the glide angle and touchdown speed are optimized, with an error ≤ 5%. For survey aircraft, buffer landing is supported when carrying heavy equipment, and the impact force is ≤ 5G.

[0025] It should be added that the same system is compatible with the differential control requirements of rotorcraft hovering landing and fixed-wing aircraft runway landing.

[0026] AI Dynamic Flight Path Generation Input data: UAV type, real-time wind speed (0 - 20 m / s), slope of takeoff and landing point (≤ 15°), obstacle distribution; Algorithm framework: a reinforcement learning model based on PPO (Proximal Policy Optimization), and the training data set contains more than 100,000 simulated landing scenarios; Output instructions: landing flight path point sequence, recommended airspeed, pitch / roll angle threshold, emergency hover instruction.

[0027] In step S4: Real-time meteorological information includes wind speed, precipitation, and visibility.

[0028] In step S5: The encrypted data link supports dual-band redundant communication (L band + C band). In an electromagnetic interference environment, the communication interruption rate is ≤ 0.1%. The flight control system of the UAV integrates on-board vision (such as infrared cameras, lidar) and ground instructions, and generates the final control signal through the Extended Kalman Filter (EKF) to perform autonomous landing operations. At the same time, for gust disturbance situations, a feedforward-feedback composite control strategy is adopted to improve the attitude stability by more than 30%.

[0029] Figure 2The present invention provides an AI-based intelligent landing guidance system for drones, including a long-distance sensing module, a laser identification and ranging module, an AI control center module, and a drone-side fusion controller module; The long-distance sensing module integrates a Ku-band phased array radar with a detection range of 5 - 10 kilometers, and a wide-angle optoelectronic tracking device for remotely detecting the position and status of drones; The laser identification and ranging module includes a 532nm laser emitter, a high-sensitivity photodetector, and a mechanical tracking platform to achieve the legality verification and high-precision ranging of drones; The AI control center module is used to deploy edge computing nodes, run a trajectory generation model based on deep reinforcement learning, integrate multi-source data, and generate dynamic landing trajectories and control instructions; The drone-side fusion controller module integrates a visual SLAM module, an instruction parsing unit, and an anti-disturbance flight control algorithm to achieve the integration of ground instructions and onboard information and control the landing process of the drone.

[0030] Experimental data Detection performance: The detection probability of MQ-9 level drones at 5 kilometers reaches 99.2%, and the positioning error is 4.3m, verifying the effectiveness of the long-distance sensing module.

[0031] Authentication test: The recognition success rate of the laser-enhanced sphere in heavy rain (rainfall of 50mm / h) is 98.5%, indicating that it can still reliably verify the identity of drones in harsh environments.

[0032] Control accuracy: For a fixed-wing aircraft (wingspan 10m) in a crosswind of 8m / s, the touchdown speed error is 2.8%, reflecting the high-precision landing control of the system for fixed-wing aircraft.

[0033] Embodiment 2 Landing of mountain area logistics drones In the mountain area logistics scenario, the phased array radar discovers a logistics rotorcraft with a load of 500kg at 5.2 kilometers, and the initial positioning error is 4.8m, obtaining its initial flight information; Use the laser target recognition module to scan the laser-enhanced sphere on the belly of the aircraft, and confirm it as a legal freight drone (code A203-B) through the dual authentication technology of nano-level reflective coating and radio frequency coding; Based on environmental information such as the real-time wind speed of 12m / s and the slope of the landing area of 3°, the AI control system generates a landing trajectory for approaching against the wind through a multi-objective optimization algorithm in combination with the type of this rotorcraft; The drone receives the trajectory instruction through an encrypted data link. Its vision system detects ground markings, and the flight control system integrates ground instructions and onboard vision information to adjust the landing attitude, and finally achieves a precise landing with a landing error of 0.12m.

[0034] Embodiment 3 Recovery of Marine Survey Aircraft On an offshore platform, a two-dimensional mechanical tracking platform tracks the roll (±6°) of a ship caused by sea waves in real time, and a laser ranging module dynamically compensates for the displacement of the deck to ensure accurate tracking of the survey aircraft; The AI model generates a landing window synchronized with the wave period based on historical sea state data at level 6 and formulates a surge-resistant landing strategy suitable for the marine environment; The survey aircraft is equipped with a landing gear with magnetorheological dampers, which effectively buffers the impact force during landing, making the landing impact force ≤4G, ensuring zero damage to the equipment and completing the precise recovery on the deck of a moving ship.

[0035] As described above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all of them should be covered by the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.

[0036] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An AI-based UAV intelligent landing guidance method, characterized in that: The steps include: Step S1, using phased array radar to perform long-distance detection of medium and large UAVs within a range of 5 kilometers, so as to obtain the initial position, speed and heading angle data of the UAV, and provide basic data for subsequent landing guidance; Step S2, scanning the laser enhanced ball on the drone through laser target recognition; Step S3, using a high-precision two-dimensional mechanical tracking platform to drive a laser rangefinder to track the three-dimensional coordinates of the UAV in real time; Step S4, integrating the terrain data of the temporary take-off and landing point, real-time weather information and the type of the UAV, deploying a multi-objective optimization algorithm, and generating a dynamic landing track and control instructions suitable for the UAV; Step S5, transmitting the generated track instruction to the UAV via an encrypted data link to perform an autonomous landing operation.

2. The AI-based UAV intelligent landing guidance method according to claim 1, characterized in that: In step S1: The phased array radar operates in the 12-18 GHz frequency band, with a positioning accuracy of ≤5m at a detection distance of 5km. It also supports simultaneous tracking of ≥20 targets, and can meet the monitoring needs of multiple drones landing at the same time.

3. The AI-based UAV intelligent landing guidance method according to claim 2 is characterized by: In step S2: The laser-enhanced ball contains a unique code and type identification, and uses nano-level reflective coating and radio frequency coding dual authentication technology. The reflectivity is ≥90% in harsh environments such as rain and fog, and the illegal drone recognition rate is ≥99.9%, accurately verifying the legality of drones and classifying them.

4. The AI-based UAV intelligent landing guidance method according to claim 3 is characterized by: In step S3: The accuracy is ≤0.1m, and its motion state can be obtained at the same time. The two-dimensional mechanical tracking platform adopts high-torque servo motor and closed-loop feedback control. The horizontal / pitch angle tracking accuracy is ≤0.01°, and the response delay is ≤10ms. It can accurately track the dynamic changes of the UAV.

5. The AI-based UAV intelligent landing guidance method according to claim 4 is characterized in that: In step S4: Different landing strategies are adopted for different types of UAVs. Rotorcraft prioritizes vertical landing accuracy with an error of ≤0.2m. Fixed-wing aircraft optimizes glide angle and touchdown speed with an error of ≤5%. Survey aircraft support cushioned landing when carrying heavy equipment, with an impact force of ≤5G.

6. The AI-based UAV intelligent landing guidance method according to claim 5 is characterized by: In step S4: Real-time weather information includes wind speed, precipitation and visibility.

7. The AI-based UAV intelligent landing guidance method according to claim 6 is characterized by: In step S5: The encrypted data link supports dual-band redundant communications. The UAV's flight control system integrates onboard vision and ground commands, generates the final control signal through extended Kalman filtering, and performs autonomous landing operations. At the same time, it adopts a feedforward-feedback composite control strategy for sudden wind disturbances.

8. An AI-based UAV intelligent landing guidance system, used to implement the AI-based UAV intelligent landing guidance method according to any one of claims 1 to 7, characterized in that: It includes long-range perception module, laser recognition and ranging module, AI control center module and drone-side fusion controller module; The long-range perception module integrates a Ku-band phased array radar and wide-angle optoelectronic tracking equipment to detect the position and status of the drone at a long distance; The laser identification and ranging module includes a laser transmitter, a high-sensitivity photoelectric detector and a mechanical tracking platform to achieve the legality verification and high-precision ranging of the drone; The AI ​​control center module is used to deploy edge computing nodes, run a trajectory generation model based on deep reinforcement learning, integrate multi-source data, and generate dynamic landing trajectories and control instructions; The drone-side fusion controller module integrates a visual SLAM module, a command parsing unit and an anti-disturbance flight control algorithm to achieve the fusion of ground commands and airborne information and control the landing process of the drone.

Images