Automatic landing system and method for boat-mounted unmanned aerial vehicle

Abstract

The invention provides an automatic landing system and method for an airship-mounted unmanned aerial vehicle, and relates to the technical field of unmanned aerial vehicles. After receiving the landing instruction, the unmanned aerial vehicle searches a nearest unmanned aerial vehicle parking apron, namely a parking apron, and advances towards the nearest parking apron; the unmanned aerial vehicle is connected with the parking apron and sends a to-be-landed instruction, and after the parking apron receives the to-be-landed instruction, self-stabilization adjustment is started, so that the parking apron is kept in a horizontal state; the unmanned aerial vehicle starts an airborne camera for target recognition, recognizes the parking apron and stably lands on the parking apron; after the unmanned aerial vehicle lands on the parking apron, the parking apron aligns and fixes the unmanned aerial vehicle, so that the unmanned aerial vehicle moves to the position of a magnetic attraction charging port in the parking apron; and the parking apron identifies the positive and negative electrodes of the charging interface of the unmanned aerial vehicle and charges the unmanned aerial vehicle. The invention aims to improve the autonomous take-off and landing capability of the unmanned aerial vehicle in a dynamic environment and expand lake surface cruising means. According to the invention, the precision of automatic landing is improved, the automation performance and cruising ability of the unmanned aerial vehicle in a dynamic and complex environment are improved, and the inspection efficiency is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to an automatic landing system and method for a boat-borne unmanned aerial vehicle. Background Art

[0002] With the popularity of quad-rotor drones in many fields, the autonomy of drone deployment in the field has also increased. In recent years, drones have developed rapidly and their related performance has been greatly improved. Despite the rapid development, there are relatively few studies on enabling drones to take off and land without human intervention. In the case of quad-rotor drones, it is impossible to land safely on inclined and uneven terrain. This is due to the physical property that the thrust of the drone is always perpendicular to the landing gear. If the drone lands on an inclined plane, it will cause the rotor to touch the ground and damage the drone structure. Therefore, a take-off and landing platform (helipad) that can always remain horizontal is necessary for the autonomous take-off and landing of drones.

[0003] In addition, the battery life issue is also a major pain point of current drone technology. The battery life issue will be more significant when drones perform tasks such as logistics distribution and patrolling, and it is a key factor restricting the efficiency of drone inspections. Although existing fuel cells or solar energy can already power drones, considering the weight and cost of such power supply systems, using lithium polymer batteries as a power source for quadcopters is the best solution available. Autonomous charging is one of the inevitable development trends of drone technology. One of the most important features of future aerial drones is full automation, that is, completing 24-hour uninterrupted tasks such as security patrols without human intervention. Therefore, autonomous charging is a necessary capability for future drones. Summary of the invention

[0004] In order to overcome the defects in the above-mentioned prior art, the present invention provides an automatic landing method for a boat-borne UAV, which improves the accuracy of automatic landing and enhances the inspection efficiency.

[0005] To achieve the above object, the present invention adopts the following technical solutions, including:

[0006] The automatic landing method of a boat-borne UAV comprises the following steps:

[0007] S1, after receiving the landing command, the UAV searches for the nearest UAV landing pad, i.e., the helipad, and moves in the direction of the nearest helipad;

[0008] S2, the drone connects to the apron and sends a landing instruction. After receiving the landing instruction, the apron starts to perform self-stabilization adjustment to keep the apron in a horizontal state;

[0009] S3, the drone turns on the onboard camera to identify the target, identifies the apron, and lands stably on the apron;

[0010] S4, after the drone lands on the apron, the apron will straighten and fix the drone, so that the drone moves to the magnetic charging port on the apron.

[0011] Preferably, after step S4, the following steps are also included:

[0012] S5, the apron identifies the positive and negative poles of the drone charging interface and charges the drone.

[0013] Preferably, in step S3, the YoloV5 target recognition algorithm is used to visually guide the landing process of the drone; two models are used for target recognition, wherein one model, namely the first model, is used to recognize the overall image of the apron, and the other model, namely the second model, is used to recognize the special character image on the apron;

[0014] The specific process of step S3 is as follows:

[0015] S31, when the height distance between the UAV and the apron exceeds the set threshold, the first model is used for target recognition, and after the apron is recognized, the UAV flies to the top of the apron;

[0016] S32, the drone then descends a set distance, and determines whether the height distance between the drone and the landing pad exceeds a set threshold. If so, continue to execute step S31, otherwise execute step S33;

[0017] S33, when the height distance between the UAV and the apron is lower than the set threshold, the second model is used to perform target recognition, and after the special characters on the apron are recognized, the UAV flies to the top of the special characters on the apron;

[0018] S34, then the drone descends a set distance to determine whether the drone can recognize the special characters on the apron; if so, continue to execute step S33, otherwise cut off the power source of the drone and land the drone.

[0019] Preferably, in step S4, a barrier bar is provided on each of the four sides of the apron, and a double-sided drive mode is adopted, and the wires are retracted and released on both sides simultaneously to achieve the effect of synchronous movement of the barrier bars.

[0020] Preferably, in step S5, a pressure-triggered lighting device is installed at the positive charging port of the drone. After the drone touches the pressure-triggered lighting device, the apron determines the light-emitting position through a photoresistor to determine the positive and negative poles of the charging port. The apron adjusts the charging polarity and connects the charging circuit.

[0021] Preferably, the specific process of step S5 is as follows:

[0022] S51, initialize the related pins of wired charging, including the light sensor pin and the relay enable pin;

[0023] S52, when it is detected that the wired charging function is enabled, the helipad is lowered to the charging position by utilizing the translational freedom of the helipad in the z direction, and the magnetic charging port contacts on the landing gear of the UAV are attracted to the magnetic contacts on the helipad;

[0024] S53, when the magnetic charging port is closed, the apron determines the light position through the photoresistor to determine the positive and negative poles of the charging port, thereby determining the polarity direction and enabling the corresponding relay to start charging;

[0025] S54, the drone is equipped with a BMS module, and the drone starts corresponding charging protection measures;

[0026] S55, after charging starts, a real-time judgment is made on whether the drone has received a take-off command. If a take-off command is received, all relays of the helipad are disabled, the power supply circuit is cut off, the drone BMS module switches to the discharge circuit, the helipad rises, and the magnetic contacts separate during the rise.

[0027] Preferably, the apron adopts a three-degree-of-freedom Stewart platform. In step S2, during the self-stabilization adjustment process, the apron needs to continuously obtain the apron attitude angle, calculate the deviation and perform correction actions to keep the apron in a horizontal state; the details are as follows:

[0028] Set a timer to generate an interrupt every Δt milliseconds. During the interrupt, first obtain the attitude angle of the apron, then calculate the control amount of each servo through PID operation, set the control amount to the desired attitude angle, and finally exit the interrupt. After the interrupt is completed, the apron uses inverse kinematics to calculate the servo angle based on the desired attitude angle and control the servo movement.

[0029] Preferably, the self-stabilization adjustment of the apron specifically includes the following steps:

[0030] S21, the self-stabilization program starts running;

[0031] S22, perform initialization settings, including: initialization of the parameters of the apron; setting the timer to generate a Δt millisecond interrupt, and calculating the RPY angle in the interrupt processing function; then returning the apron to zero, that is, setting the angle of each servo to 0, and performing IMU calibration after a delay, obtaining the values ​​of the Roll angle and Pitch angle at this time, and taking them as the zero point;

[0032] S23, perform apron configuration, including: clock, serial port and related pin configuration;

[0033] S24, generating a cyclic interrupt of Δt milliseconds through a timer;

[0034] The interruption process is as follows: read the raw sensor data of the gyroscope, calculate the corresponding Roll angle and Pitch angle respectively through the DMP library, and use the position PI regulator to adjust the Roll angle and Pitch angle;

[0035] S25, after the interruption is completed, the apron is returned to zero, the servo angle is returned to zero, and after a delay, the IMU is calibrated, the values ​​of the Roll angle and the Pitch angle are read as the zero point, and the IMU is calibrated to zero point, so that its self-balancing structure converges to the expected value;

[0036] S26, calculate the I vector according to phi-set, pitch-set and z-set, obtain the required angle of the servo through the I vector, and then set the calculated servo angle as the target servo angle, adjust the current servo angle to the target angle, and finally realize the self-stabilizing cycle process.

[0037] The present application also provides a boat-borne UAV automatic landing system, which is applied to the above-mentioned boat-borne UAV automatic landing method, and the system includes a UAV and a helipad;

[0038] The drone includes a visual recognition module, a GPS module, a wireless module and an onboard computer;

[0039] The apron includes a steering gear, an attitude sensor and a program control system;

[0040] The onboard computer on the UAV is in communication connection with the program control system of the apron.

[0041] Preferably, the helipad is arranged on the water surface, and the drones inspecting the water area can land on the helipad and be charged.

[0042] The advantages of the present invention are:

[0043] (1) The collaborative mode of UAV and apron can innovatively combine the inspection methods of air and water, forming an efficient joint operation mechanism. That is, combining the aerial vision advantage of UAV and the stable navigation characteristics of apron, it can achieve full coverage of the target area and improve operation efficiency.

[0044] (2) The autonomous charging design reduces mission interruptions or delays caused by insufficient power, thereby improving inspection efficiency and solving the two core challenges of charging port alignment and positive and negative pole switching after the drone lands.

[0045] (3) The helipad adopts the Stewart platform technology and combines it with advanced control algorithms to provide a highly stable guarantee for the landing of drones. The improved Stewart platform has multi-degree-of-freedom movement capabilities to ensure that the platform always maintains a horizontal posture in different environments. The integrated control algorithm can respond quickly and adjust the platform angle in real time. The helipad has an attitude sensor that can identify the attitude information of the helipad top plate; it is necessary to have an actuator that can adjust the helipad attitude to reduce the interference of wind and waves on the helipad, and try to keep the helipad plane in a horizontal posture that is suitable for drones to dock.

[0046] (4) The drone landing uses a dual positioning method that combines GPS and vision to ensure accurate navigation. GPS positioning provides accurate data on airspace location, while YOLO visual detection technology based on deep learning analyzes images in real time to identify and lock the precise location of the apron. Through the synergy of these two positioning methods, the drone can continuously obtain high-precision location information during the landing process, ensuring the safety and accuracy of the landing process. Two models are used for different target recognition to ensure that the drone is always in a closed-loop state of visual positioning during the landing process, and blind landing will not occur, which can increase the probability of accurate landing.

[0047] (5) When the UAV lands on the helipad with any nose orientation, the return mechanism can adjust the UAV to a specific position and play a certain fixing role at the same time; and it has an electromagnetic attraction module, which can play a certain auxiliary role in the landing of the UAV and can also play a role in fixing the UAV to prevent the boat-borne UAV from detaching from the helipad surface and causing damage due to wind and wave interference; it has a positive and negative pole identification module, which can identify the positive and negative pole directions of the UAV charging interface, automatically switch the positive and negative pole directions to complete the wired charging of the UAV, and improve the UAV's endurance. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The figure is a flow chart of the automatic landing method of the boat-borne UAV of the present invention.

[0049] Figure 2 The figure is a flow chart of the self-stabilizing adjustment of the apron of the present invention.

[0050] Figure 3 Schematic diagram of the coordinate system from the base to the platform of the three-degree-of-freedom Stewart platform.

[0051] Figure 4 It is a schematic diagram of the return mechanism of the helipad of the present invention.

[0052] Figure 5 Schematic diagram of the drone of the present invention.

[0053] Figure 6It is a schematic diagram of the helipad of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 creative work are within the scope of protection of the present invention.

[0055] The automatic landing system of a boat-borne UAV includes a UAV and a helipad. The UAV mainly includes a visual recognition module, a GPS module, a wireless module and an onboard computer. The helipad mainly includes a steering gear, a sensor and a program control system. The onboard computer on the UAV is connected to the program control system of the helipad in communication.

[0056] The helipad can be set up on the water surface, and the drone can stop on the helipad to charge when conducting space inspections over water areas.

[0057] Depend on Figure 1 As shown, the automatic landing method of the boat-borne UAV includes the following steps:

[0058] S1, the drone receives a landing command due to its own reasons or external commands; the drone searches for the nearest helipad (self-stabilizing platform) and obtains the GPS coordinates of the helipad. If the distance is far, the drone first uses GPS for navigation to keep the drone moving steadily towards the helipad.

[0059] S2, the drone uses a wireless module to connect to the apron and sends a landing instruction. After receiving the landing instruction, the apron starts to adjust itself. The apron needs to continuously obtain the apron attitude angle, calculate the deviation and perform correction actions to keep the apron horizontal. In the present invention, the apron uses a three-degree-of-freedom Stewart platform.

[0060] Depend on Figure 2 As shown in FIG. 1 , the apron self-stabilization adjustment specifically includes the following steps:

[0061] S21, the self-stabilization program starts running.

[0062] S22, perform a series of initialization settings:

[0063] Initialization of apron parameters: mainly to initialize various known conditions of the apron, such as the length of the connecting rod;

[0064] Initialization of the general timer: generate a 5ms interrupt and calculate the RPY angle in the interrupt processing function;

[0065] Then reset the apron to zero, that is, set the angle of each servo to 0, and perform IMU calibration after a delay, obtain the values ​​of the Roll angle and Pitch angle at this time, and use them as the zero point.

[0066] S23, performs a series of apron configurations, such as clock, serial port and related pin configurations.

[0067] S24, generates a 5ms cyclic interrupt through the timer. During the interrupt, the following operations are performed:

[0068] Read the raw sensor data of the gyroscope (model: MPU6050), and then calculate the corresponding roll angle and pitch angle respectively through the DMP library, and use the position PI regulator to adjust the roll angle and pitch angle.

[0069] S25, after completing the above interruption, the apron is zeroed, the servo angle is zeroed, and after a delay of 5s, the IMU is calibrated, the values ​​of Roll and Pitch are read as the zero point, and the IMU is calibrated to zero point so that its self-balancing structure converges to the expected value.

[0070] S26, calculate the I vector according to phi-set, pitch-set and z-set, obtain the required angle of the servo through the I vector, and then set the calculated servo angle as the target servo angle, and adjust the current servo angle to the target servo angle through adjustment, finally realizing the cycle process of self-stabilization adjustment.

[0071] In order to realize self-stabilizing control, the present invention needs to continuously obtain the attitude angle of the apron, calculate the deviation and perform the correction action. For this purpose, a general timer is set for interruption, and the selected timer is TIM2, which generates an interrupt every 5ms. During the interruption process, the attitude angle of the apron is first obtained, and then the control amount of each servo is calculated through PID operation, and the control amount is set to the expected attitude angle, and finally the interruption is exited; after the interruption is completed, the main loop of the apron program will use inverse kinematics solution according to the set expected attitude angle, calculate the servo angle, and control the servo movement.

[0072] The present invention sets an adjustment dead zone near the zero point, that is, when the error is detected to be less than a certain value near the zero point, it will be considered as an error caused by the zero drift of the sensor, and no adjustment will be performed. After adopting the method of setting the adjustment dead zone near the zero point, the phenomenon of apron vibration is suppressed, and normal operation can be performed when the error is greater than the dead zone value.

[0073] Depend on Figure 3 As shown, for the i-th leg of the apron, i.e., the i-th actuator, define P i is the anchor point of the actuator on the apron, B iis the anchor point of the actuator on the base, T is the translation vector from the base (bottom plate) to the platform (top plate), and the rotation matrix from the bottom plate to the top plate is P i Vector q relative to the origin of the base coordinate system i for

[0074] The three-degree-of-freedom Stuart platform has three degrees of freedom relative to the base plate, namely up and down motion along the z-axis and rotational motion around the x-axis and y-axis; when the apron is in the initial position, z=0, and upward motion is the positive direction; the rotation angle around the x-axis is Rotate the angle θ around the y-axis and the angle ψ around the z-axis. Use the formula to get the rotation matrix from the bottom plate to the top plate from the three known quantities.

[0075]

[0076] Calculate the angle β i , angle φ i :

[0077]

[0078] Among them, b i is the vector from the origin of the base to the edge of the base; l i is the vector representation from the edge of the base to the edge of the platform; h1 and h2 are the lengths of the two connecting rods, which are known quantities.

[0079] Calculate the angle α that the servo needs to rotate i =β i -φ i .

[0080] It is difficult to solve the forward kinematics of the built parallel robot, but the inverse kinematics does not need to be solved through forward kinematics and can be directly solved through geometric methods. Therefore, the inverse kinematics is directly solved in kinematic analysis.

[0081] It is known that the two coordinate systems only rotate ψ relative to the z-axis, so the following mathematical derivation can be obtained:

[0082] x=x'cosψ-y'sinψ;

[0083] y=x'sinψ+y'cosψ;

[0084] z=z';

[0085] Therefore, we can combine the above formula to define the rotation matrix R around the z axis z (ψ) is:

[0086]

[0087] in,

[0088] Similarly, we can derive the rotation matrix R around the y-axis y (θ) and the rotation matrix around the x-axis They are:

[0089]

[0090] Following the order of rotation of the ZYX axes, the complete rotation matrix of the apron platform relative to the base is:

[0091]

[0092]

[0093] S3, the drone turns on the onboard camera to identify the target, identifies the apron, and lands steadily on the apron.

[0094] The present invention adopts YoloV5 target recognition algorithm to visually guide the landing process of the drone. Two models are separately made during the visual training process, one of which is used to identify the overall image of the apron, and the other is used to identify the H character image on the apron. Due to the limitation of camera physical factors, the apron cannot be identified as a whole at a height of about 68cm from the apron. When the distance is less than the limit distance, the H character image is used to achieve positioning. In this way, it can be ensured that the drone is always in a state of visual positioning closed loop during the landing process, and blind landing will not occur. When stable recognition can be achieved, the drone descends 10cm each time and adjusts the posture in time. When the H character image cannot be stably recognized, the drone power source is cut off, the motor stops and the landing is performed.

[0095] The specific process is as follows:

[0096] S31, when the height distance between the UAV and the apron exceeds 68 cm, the first model is used for target recognition, and after the apron is recognized, the UAV flies to the top of the apron;

[0097] S32, then the drone descends 10 cm, and it is determined whether the height distance between the drone and the landing pad exceeds 68 cm. If so, continue to step S31, otherwise, execute step S33;

[0098] S33, when the height distance between the UAV and the apron is less than 68 cm, the second model is used for target recognition, and after the H character on the apron is recognized, the UAV flies to the top of the special character on the apron;

[0099] S34, then the drone descends 10cm to determine whether the drone can stably recognize the special characters on the apron; if so, continue to execute step S33, otherwise cut off the power source of the drone and land the drone.

[0100] S4, after the drone lands on the apron, the apron will straighten and fix the drone, so that the drone moves to the magnetic charging port on the apron.

[0101] Depend on Figure 4 As shown, a barrier is set on each of the four sides of the apron, and a double-sided drive method is adopted. The wires are retracted and released on both sides at the same time to achieve the effect of synchronous movement of the barrier. Each axis requires 6 steel wire ropes, 4 of which are used to connect the active anchor points on the optical axis, and 2 are used to connect other steel wire ropes and winding wheels. The two sides are used for extension and the two sides are used for contraction to drive the active anchor closer to the center. After the return function is completed, the drone has been fixed at this time, and the apron self-stabilization function is turned off.

[0102] S5, the apron identifies the positive and negative poles of the drone charging interface and charges the drone through the apron.

[0103] Install a pressure-triggered light device next to the positive charging port of the drone to initialize the wired charging related pins, mainly the light sensor pin and the relay enable pin. First, wait for 10 minutes to cool down the drone and determine whether the charging system is enabled. When it is determined that the drone touches the light-emitting diode, the light-sensitive resistor is used to determine the positive and negative poles of the charging, adjust the charging polarity and connect the charging circuit.

[0104] In step S5, the charging steps are as follows:

[0105] S51, initialize the wired charging related pins, mainly the photosensor pin and the relay enable pin.

[0106] S52, when it is detected that the wired charging function is enabled, the translational freedom of the apron in the z direction is utilized, and after the apron is lowered to the charging position, the magnetic suction contacts on the landing gear of the drone will be attracted to the magnetic suction contacts on the apron.

[0107] S53, when the magnetic charging port is attracted, the apron determines the light position through the photoresistor to determine the positive and negative poles of the charging port, thereby determining the polarity direction and enabling the corresponding relay to start charging.

[0108] The pressure-triggered lighting device on the drone's take-off and landing will illuminate a certain photosensor. The threshold of the digital quantity output by the photoresistor can be adjusted by adjusting the potentiometer on the sensor. After adjusting to the appropriate position, the white light emitted by the LED on the pressure-triggered lighting device can be identified to determine the polarity direction and enable the corresponding relay to start charging according to the situation.

[0109] S54, the drone is equipped with a balanced charging and discharging module (BMS module), and has corresponding protection measures for this.

[0110] The BMS module has 3 ports, connected to the aircraft battery, load and charging port respectively. The GND of the above three ports are connected together. The relay has 5 contacts, namely the normally open contact, normally closed contact, input contact and the positive and negative contacts of the relay coil. Connect the positive pole of the load to the normally closed contact and the aircraft battery to the input contact, and the discharge can be realized by default; connect the positive pole of the charging port to the normally open contact, and the other one to the positive pole of the relay coil after the LDO is reduced to 12V, so that when the voltage of the charging port is above 16V, the relay will be activated and switched to the charging circuit to start charging; when there is no voltage at the charging port, the relay is in the default state of the discharge circuit.

[0111] S55, after charging starts, it will determine whether a take-off command has been received. If a take-off command is received, the helipad will first disable all relays and cut off the power supply circuit. The drone BMS module will automatically switch to the discharge circuit and then make the helipad rise. During the rise, the magnetic contacts will separate automatically and will not interfere with the drone's take-off.

[0112] S6: The drone receives the take-off command, the charging function is turned off on the helipad, and the self-stabilization adjustment is turned on to make the drone on the helipad return to its original position, and then the fixation is released and the drone takes off.

[0113] Depend on Figure 5 and Figure 6 As shown, it is a schematic diagram of the UAV and the helipad of this embodiment. The UAV is provided with a monocular camera 11, a propeller 12, a flight control module 13, an onboard computer 14 and a UAV charging interface 15; the helipad is provided with a barrier bar 21, a helipad charging interface 22, a hangar power distribution board 23 and a servo 24.

[0114] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for automatically landing a boat-borne UAV, characterized in that: The following steps are involved: S1, after receiving the landing command, the UAV searches for the nearest UAV landing pad, i.e., the helipad, and moves in the direction of the nearest helipad; S2, the drone connects to the apron and sends a landing instruction. After receiving the landing instruction, the apron starts to perform self-stabilization adjustment to keep the apron in a horizontal state; S3, the drone turns on the onboard camera to identify the target, identifies the apron, and lands stably on the apron; S4, after the drone lands on the apron, the apron will straighten and fix the drone, so that the drone moves to the magnetic charging port on the apron.

2. The automatic landing method of a boat-borne UAV according to claim 1, characterized in that: The step S4 further includes the following steps: S5, the apron identifies the positive and negative poles of the drone charging interface and charges the drone.

3. The automatic landing method of a boat-borne UAV according to claim 1, characterized in that: In step S3, the YoloV5 target recognition algorithm is used to visually guide the landing process of the drone; two models are used for target recognition, one model, namely the first model, is used to recognize the overall image of the apron, and the other model, namely the second model, is used to recognize the special character image on the apron; The specific process of step S3 is as follows: S31, when the height distance between the UAV and the apron exceeds the set threshold, the first model is used for target recognition, and after the apron is recognized, the UAV flies to the top of the apron; S32, the drone then descends a set distance, and determines whether the height distance between the drone and the landing pad exceeds a set threshold. If so, continue to execute step S31, otherwise execute step S33; S33, when the height distance between the UAV and the apron is lower than the set threshold, the second model is used to perform target recognition, and after the special characters on the apron are recognized, the UAV flies to the top of the special characters on the apron; S34, then the drone descends a set distance to determine whether the drone can recognize the special characters on the apron; if so, continue to execute step S33, otherwise cut off the power source of the drone and land the drone.

4. The automatic landing method of a boat-borne UAV according to claim 1, characterized in that: In step S4, a barrier bar is provided on each of the four sides of the apron, and a double-sided drive mode is adopted, and the wires are retracted and released on both sides simultaneously to achieve the effect of synchronous movement of the barrier bars.

5. The automatic landing method of a boat-borne UAV according to claim 2, characterized in that: In step S5, a pressure-triggered lighting device is installed at the positive charging port of the drone. When the drone touches the pressure-triggered lighting device, the apron determines the light-emitting position through the photoresistor to determine the positive and negative poles of the charging port. The apron adjusts the charging polarity and connects the charging circuit.

6. The automatic landing method of a boat-borne UAV according to claim 5, characterized in that: The specific process of step S5 is as follows: S51, initialize the related pins of wired charging, including the light sensor pin and the relay enable pin; S52, when it is detected that the wired charging function is enabled, the helipad is lowered to the charging position by utilizing the translational freedom of the helipad in the z direction, and the magnetic charging port contacts on the landing gear of the UAV are attracted to the magnetic contacts on the helipad; S53, when the magnetic charging port is closed, the apron determines the light position through the photoresistor to determine the positive and negative poles of the charging port, thereby determining the polarity direction and enabling the corresponding relay to start charging; S54, the drone is equipped with a BMS module, and the drone starts corresponding charging protection measures; S55, after charging starts, a real-time judgment is made on whether the drone has received a take-off command. If a take-off command is received, all relays of the helipad are disabled, the power supply circuit is cut off, the drone BMS module switches to the discharge circuit, the helipad rises, and the magnetic contacts separate during the rise.

7. The automatic landing method of a boat-borne UAV according to claim 1, characterized in that: The apron adopts a three-degree-of-freedom Stewart platform. In step S2, during the self-stabilization adjustment process, the apron needs to continuously obtain the apron attitude angle, calculate the deviation and perform correction actions to keep the apron in a horizontal state; the details are as follows: Set a timer to generate an interrupt every Δt milliseconds. During the interrupt, first obtain the attitude angle of the apron, then calculate the control amount of each servo through PID operation, set the control amount to the desired attitude angle, and finally exit the interrupt; After the interruption is completed, the apron uses inverse kinematics to calculate the servo angle based on the desired attitude angle and control the servo movement.

8. The automatic landing method of a boat-borne UAV according to claim 2, characterized in that: The self-stabilization adjustment of the apron specifically includes the following steps: S21, the self-stabilization program starts running; S22, perform initialization settings, including: initialization of the parameters of the apron; setting the timer to generate a Δt millisecond interrupt, and calculating the RPY angle in the interrupt processing function; then returning the apron to zero, that is, setting the angle of each servo to 0, and performing IMU calibration after a delay, obtaining the values ​​of the Roll angle and Pitch angle at this time, and taking them as the zero point; S23, perform apron configuration, including: clock, serial port and related pin configuration; S24, generating a cyclic interrupt of Δt milliseconds through a timer; The interruption process is as follows: read the raw sensor data of the gyroscope, calculate the corresponding Roll angle and Pitch angle respectively through the DMP library, and use the position PI regulator to adjust the Roll angle and Pitch angle; S25, after the interruption is completed, the apron is returned to zero, the servo angle is returned to zero, and after a delay, the IMU is calibrated, the values ​​of the Roll angle and the Pitch angle are read as the zero point, and the IMU is calibrated to zero point, so that its self-balancing structure converges to the expected value; S26, calculate the I vector according to phi-set, pitch-set and z-set, obtain the required angle of the servo through the I vector, and then set the calculated servo angle as the target servo angle, adjust the current servo angle to the target angle, and finally realize the self-stabilizing cycle process.

9. The automatic landing system of a boat-borne UAV is characterized by: The automatic landing method for a boat-borne UAV as described in any one of claims 1 to 9 above, wherein the system comprises a UAV and a helipad; The drone includes a visual recognition module, a GPS module, a wireless module and an onboard computer; The apron includes a steering gear, an attitude sensor and a program control system; The onboard computer on the UAV is in communication connection with the program control system of the apron.

10. The automatic landing system for a boat-borne UAV according to claim 9, characterized in that: The helipad is arranged on the water surface, and the drones inspecting the water area can land on the helipad and be charged.

Images