Unmanned aerial vehicle fixed-point landing control method, equipment and medium

Abstract

The embodiment of the invention discloses an unmanned aerial vehicle fixed-point landing control method and device and a medium, belongs to the technical field of unmanned aerial vehicles, and solves the problem that the fixed-point landing deviation of the unmanned aerial vehicle is large in a complex environment and under the high-precision landing requirement. Obtaining the relative position between the unmanned aerial vehicle and the target landing area, dividing the landing space into a plurality of landing layers, and determining the height difference between the landing layers; based on the height difference value, the current flight parameter corresponding to the unmanned aerial vehicle and the wind field parameter corresponding to each landing layer, constructing a prediction trajectory corresponding to each landing layer; splicing the prediction trajectories corresponding to the landing layers to obtain an unmanned aerial vehicle landing trajectory; acquiring a wind field parameter corresponding to the target landing area, and determining a reference landing area based on the wind field parameter and the prediction trajectory; and based on the position deviation between the target landing area and the reference landing area, adjusting the landing trajectory of the unmanned aerial vehicle to control the unmanned aerial vehicle to land to the target landing area.

Description

Technical Field

[0001] The present application relates to the field of UAV technology, and in particular to a method, device and medium for controlling the fixed-point landing of an UAV. Background Art

[0002] With the rapid development of drone technology, it has been widely used in many fields, such as logistics distribution, agricultural plant protection, surveying and exploration, and emergency rescue. In these application scenarios, the precise landing of drones is crucial, which is directly related to the successful execution of the mission and the safe recovery of the equipment.

[0003] In recent years, some advanced sensor technologies, such as LiDAR and visual sensors, have begun to be applied to the landing process of drones. LiDAR can measure the distance information of the surrounding environment by emitting laser beams, construct a three-dimensional point cloud map, and assist drones in precise positioning. However, LiDAR is expensive, which to a certain extent limits its large-scale application. In addition, under severe weather conditions, such as strong winds, heavy rain, and dense fog, the laser beam will be scattered and absorbed by raindrops and fog particles, resulting in a significant shortening of the measurement distance and a serious decrease in data accuracy.

[0004] Therefore, the existing UAV fixed-point landing method has large errors when facing complex environments and high-precision landing requirements, resulting in a large deviation from the target landing area. Summary of the invention

[0005] The embodiments of the present application provide a method, device and medium for controlling the fixed-point landing of a UAV, which are used to solve the following technical problems: When facing complex environments and high-precision landing requirements, the existing fixed-point landing method of a UAV has large errors, resulting in a large deviation from the target landing area.

[0006] The present application embodiment adopts the following technical solutions:

[0007] An embodiment of the present application provides a method for controlling a drone to land at a fixed point. The method includes: when receiving a drone landing request, obtaining the relative position between the drone and the target landing area; based on the relative position, dividing the landing space into multiple landing layers and determining the height difference between each landing layer; based on the height difference, the current flight parameters corresponding to the drone, and the wind field parameters corresponding to each landing layer respectively, constructing the predicted trajectories corresponding to each landing layer respectively; splicing the predicted trajectories corresponding to each landing layer respectively to obtain the drone landing trajectory; obtaining the wind field parameters corresponding to the target landing area, and based on the wind field parameters and the predicted trajectory, determining the reference landing area; determining the position deviation between the target landing area and the reference landing area, and based on the position deviation, determining the flight parameters to be adjusted for the drone, so as to adjust the drone landing trajectory based on the flight parameters to be adjusted for the drone; and controlling the drone to land at the target landing area through the adjusted drone landing trajectory.

[0008] In an embodiment of the present application, the landing layers are divided based on the relative position between the drone and the target landing area, fully considering the environmental differences at different heights. The determination of the height difference between each landing layer enables the drone to perform segmented trajectory planning according to the actual situation. The predicted trajectories are constructed by comprehensively considering the height difference, the current flight parameters, and the wind field parameters of each layer, improving the accuracy of the predicted trajectories of each landing layer. By comparing the position deviation between the target landing area and the reference landing area and dynamically adjusting the flight parameters according to the position deviation, the drone can quickly adapt to environmental changes, always maintain the trend of flying towards the target landing area, effectively reduce the distance deviation from the target area, and improve the landing adaptability and accuracy of the drone in a complex and changeable environment. Determining the reference landing area based on the wind field parameters and the predicted trajectory provides a more targeted reference range for the drone to land, thereby improving the success rate and safety of landing.

[0009] In an implementation manner of the present application, constructing the predicted trajectories corresponding to each landing layer respectively based on the height difference, the current flight parameters corresponding to the drone, and the wind field parameters corresponding to each landing layer respectively specifically includes: inputting the current flight parameters, the height difference, and the wind field parameters corresponding to the current layer of the drone into a preset trajectory prediction model, so as to output the predicted trajectory of the drone landing from the current position to the next layer through the trajectory prediction model; obtaining the predicted area based on the predicted trajectory and determining the area deviation between the predicted area and the preset area range; in the case that the area deviation does not meet the preset conditions, inputting the area deviation and the wind field parameters corresponding to the current layer of the drone into a preset flight parameter adjustment model, so as to output the flight adjustment parameters corresponding to the current layer of the drone through the preset flight parameter adjustment model; adjusting the flight parameters of the drone based on the flight adjustment parameters corresponding to the current layer, and re-outputting the predicted trajectory of the drone landing from the current position to the next layer through the adjusted flight parameters.

[0010] In an implementation manner of the present application, a prediction area is obtained based on a predicted trajectory, which specifically includes: in the predicted trajectory, after a preset time interval, trajectory points are screened, and a set of trajectory points is constructed based on the screened trajectory points; the set of trajectory points is classified based on spatial positions; based on the classified set of trajectory points, a boundary contour of the predicted trajectory in space is constructed; based on the highest point and the lowest point in the boundary contour, the boundary range in the vertical direction is determined; and, the minimum convex polygon containing all trajectory points is determined through a convex hull algorithm, and the boundary of the minimum convex polygon is used as the boundary range of the predicted trajectory in the horizontal direction; based on the boundary range in the vertical direction and the boundary range in the horizontal direction, a prediction range is obtained.

[0011] In an implementation manner of the present application, the predicted trajectories corresponding to each landing layer are spliced to obtain a UAV landing trajectory, which specifically includes: determining the position deviation between the trajectory starting point of each landing layer and the center point of the target landing area, and determining the direction deviation between adjacent predicted trajectories; based on the position deviation and the direction deviation, the predicted trajectories of each landing layer are translated and rotated in the space coordinate system so that the starting point and the ending point of the UAV in each landing layer are connected; based on the connected predicted trajectories, Bezier curves are constructed in each landing layer to generate transition trajectories at each connection point through the Bezier curves; the adjusted predicted trajectories and the transition trajectories corresponding to each landing layer are spliced to obtain a UAV landing trajectory.

[0012] In an implementation manner of the present application, wind field parameters corresponding to a target landing area are obtained, and a reference landing area is determined based on the wind field parameters and the predicted trajectory, which specifically includes: extracting features of the predicted trajectory; where the extracted features at least include one of the trajectory curvature change value, the speed change trend, and the spatial distribution of trajectory points at different time periods; based on each time step of the predicted trajectory, according to the wind field parameters at the current position of the UAV, the position data of the UAV at the next moment is generated through Monte Carlo simulation; the generated position data at the next moment is classified through a vector machine, and the position data at the next moment falling within the safe range of the target landing area is marked as positive class data, and other position data at the next moment is marked as negative class data; based on the positive class data, a reference landing trajectory is determined.

[0013] In an implementation manner of the present application, the position deviation between the target landing area and the reference landing area is determined, and based on the position deviation, the flight parameters to be adjusted for the unmanned aerial vehicle are determined, specifically including: constructing a three-dimensional space coordinate system based on the target landing area and the reference landing area; performing weight allocation on the pseudorange measured by the GNSS of the unmanned aerial vehicle and the differential correction data of RTK obtained based on the flight state of the unmanned aerial vehicle and environmental information; determining the position deviation in the three-dimensional space coordinate system through the data after weight allocation; where the position deviation includes translational deviation and rotational deviation; inputting the position deviation, the current pose of the unmanned aerial vehicle, and the wind field parameters corresponding to the target landing area into a preset target parameter adjustment model to output the flight parameters to be adjusted for the unmanned aerial vehicle through the preset target parameter adjustment model.

[0014] In an implementation manner of the present application, based on the flight state of the unmanned aerial vehicle and environmental information, weight allocation is performed on the pseudorange measured by the GNSS of the unmanned aerial vehicle and the differential correction data of RTK, specifically including:

[0015] Through the function:

[0016]

[0017] Determine the weight of the pseudorange measured by GNSS; based on the function:

[0018] W RTK = 1 - W GNSS ;

[0019] Determine the weight of the RTK differential correction data; where W GNSS is the weight of the pseudorange measured by GNSS; S GNSS is the GNSS signal strength; S RTK is the RTK signal strength; γ is the first adjustment coefficient; V is the flight speed; V max is the maximum flight speed of the unmanned aerial vehicle; δ is the second adjustment coefficient; H is the flight height; H max is the maximum flight height of the unmanned aerial vehicle; ε is the third adjustment coefficient; σ is the standard deviation of terrain undulation; β is the fourth adjustment coefficient; α is the fifth adjustment coefficient; W RTK is the weight of the RTK differential correction data.

[0020] In an implementation manner of the present application, the landing trajectory of the unmanned aerial vehicle is adjusted based on the flight parameters to be adjusted for the unmanned aerial vehicle, specifically including: determining the amount to be adjusted corresponding to the next trajectory point based on the flight parameters to be adjusted, so as to re-plan the coordinates of the next trajectory point based on the amount to be adjusted; constructing trajectory discrete points based on the adjusted next trajectory points, and connecting the trajectory discrete points into a curve through a spline curve fitting algorithm; obtaining the adjusted landing trajectory of the unmanned aerial vehicle based on the connected curve. W RTK is the weight of the RTK differential correction data.

[0021] An embodiment of the present application provides a fixed-point landing control device for an unmanned aerial vehicle, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: when receiving a landing request of the unmanned aerial vehicle, obtain the relative position between the unmanned aerial vehicle and the target landing area, divide the landing space into multiple landing layers based on the relative position, and determine the height difference between the landing layers; construct prediction trajectories corresponding to the respective landing layers based on the height difference, the current flight parameters corresponding to the unmanned aerial vehicle, and the wind field parameters corresponding to the respective landing layers; splice the prediction trajectories corresponding to the respective landing layers to obtain a landing trajectory of the unmanned aerial vehicle; obtain the wind field parameters corresponding to the target landing area, and determine a reference landing area based on the wind field parameters and the prediction trajectory; determine the position deviation between the target landing area and the reference landing area, and determine the flight parameters to be adjusted for the unmanned aerial vehicle based on the position deviation, so as to adjust the landing trajectory of the unmanned aerial vehicle based on the flight parameters to be adjusted for the unmanned aerial vehicle; control the unmanned aerial vehicle to land on the target landing area through the adjusted landing trajectory of the unmanned aerial vehicle.

[0022] A non-volatile computer storage medium provided by an embodiment of the present application stores computer-executable instructions, and the computer-executable instructions are set to: when receiving a landing request of the unmanned aerial vehicle, obtain the relative position between the unmanned aerial vehicle and the target landing area, divide the landing space into multiple landing layers based on the relative position, and determine the height difference between the landing layers; construct prediction trajectories corresponding to the respective landing layers based on the height difference, the current flight parameters corresponding to the unmanned aerial vehicle, and the wind field parameters corresponding to the respective landing layers; splice the prediction trajectories corresponding to the respective landing layers to obtain a landing trajectory of the unmanned aerial vehicle; obtain the wind field parameters corresponding to the target landing area, and determine a reference landing area based on the wind field parameters and the prediction trajectory; determine the position deviation between the target landing area and the reference landing area, and determine the flight parameters to be adjusted for the unmanned aerial vehicle based on the position deviation, so as to adjust the landing trajectory of the unmanned aerial vehicle based on the flight parameters to be adjusted for the unmanned aerial vehicle; control the unmanned aerial vehicle to land on the target landing area through the adjusted landing trajectory of the unmanned aerial vehicle.

[0023] The above at least one technical solution adopted in the embodiments of the present application can achieve the following beneficial effects: The embodiments of the present application divide the landing layers based on the relative position between the unmanned aerial vehicle (UAV) and the target landing area, fully considering the environmental differences at different heights. The determination of the height difference between each landing layer enables the UAV to perform segmented trajectory planning according to the actual situation. By comprehensively considering the height difference, the current flight parameters, and the wind field parameters of each layer, a predicted trajectory is constructed, improving the accuracy of the predicted trajectory for each landing layer. By comparing the position deviation between the target landing area and the reference landing area and dynamically adjusting the flight parameters according to the position deviation, the UAV can quickly adapt to environmental changes, always maintain the trend of flying towards the target landing area, effectively reduce the distance deviation from the target area, and improve the landing adaptability and accuracy of the UAV in complex and changeable environments. Determining the reference landing area based on the wind field parameters and the predicted trajectory provides a more targeted reference range for the UAV to land, thereby improving the success rate and safety of landing. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0025] Figure 1 It is a flowchart of a method for controlling a UAV to land at a fixed point provided by an embodiment of the present application;

[0026] Figure 2 It is a schematic structural diagram of a device for controlling a UAV to land at a fixed point provided by an embodiment of the present application.

[0027] Reference Numerals:

[0028] 200: Device for controlling a UAV to land at a fixed point, 201: Processor, 202: Memory. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] The embodiments of the present application provide a method, device, and medium for controlling a UAV to land at a fixed point.

[0030] In order to enable those skilled in the art to better understand the technical solutions in the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0031] The technical solutions proposed in the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0032] Figure 1 It is a flowchart of a method for controlling a drone to land at a fixed point provided by an embodiment of the present application. As Figure 1 shown, the method for controlling a drone to land at a fixed point includes the following steps:

[0033] S101. When a drone landing request is received, obtain the relative position between the drone and the target landing area. Based on the relative position, divide the landing space into multiple landing layers, and determine the height difference between each landing layer.

[0034] In one implementation manner of the present application, when the control system of the drone receives a landing request, the relative position relationship between the drone and the target landing area is determined through sensors and a positioning system. Based on this relative position information, the entire landing space from the current position of the drone to the target landing area is divided into multiple different landing layers in the vertical direction. The height of the landing layer will vary based on environmental conditions and flight control requirements at different heights.

[0035] For example, in the low-altitude area close to the ground, there are more obstacles, and the airflow may be more complex and unstable; while in the higher airspace, although the wind field may be relatively stable, the distance to the target landing area is far, and the flight attitude and speed need to be adjusted gradually. By stratifying the landing space, appropriate flight strategies and trajectory planning can be formulated for each layer according to its characteristics. The number of layers divided can be adjusted according to the actual situation. For example, when the landing space is large and the environment is complex, more layers may be divided to achieve more refined control; while in a simple environment, the number of layers can be relatively small. After completing the stratification of the landing space, the height difference between each landing layer needs to be determined.

[0036] S102. Based on the height difference, the current flight parameters corresponding to the drone, and the wind field parameters corresponding to each landing layer, construct the predicted trajectory corresponding to each landing layer.

[0037] In an implementation manner of the present application, the current flight parameters, the altitude difference, and the wind field parameters corresponding to the current layer of the unmanned aerial vehicle are input into a preset trajectory prediction model, so as to output, through the trajectory prediction model, a predicted trajectory of the unmanned aerial vehicle landing from the current position to the next layer. Based on the predicted trajectory, a predicted area is obtained, and an area deviation between the predicted area and the preset area range is determined. In the case that the area deviation does not meet the preset conditions, the area deviation and the wind field parameters corresponding to the current layer of the unmanned aerial vehicle are input into a preset flight parameter adjustment model, so as to output, through the preset flight parameter adjustment model, the flight adjustment parameters corresponding to the current layer of the unmanned aerial vehicle. Based on the flight adjustment parameters corresponding to the current layer, the flight parameters of the unmanned aerial vehicle are adjusted, and through the adjusted flight parameters, the predicted trajectory of the unmanned aerial vehicle landing from the current position to the next layer is output again.

[0038] Specifically, when the unmanned aerial vehicle is in the landing process, first, the current flight parameters, the altitude difference, and the wind field parameters corresponding to the current layer of the unmanned aerial vehicle are input into the preset trajectory prediction model. The current flight parameters include information such as the speed, direction, and attitude of the unmanned aerial vehicle, and these parameters reflect the current flight state of the unmanned aerial vehicle. The altitude difference refers to the vertical distance between the current landing layer and the next layer. The wind field parameters describe meteorological conditions such as the wind force magnitude and wind direction at the current altitude layer. The predicted trajectory output by the preset trajectory prediction model is a path in three-dimensional space, showing the possible flight route of the unmanned aerial vehicle under the influence of various factors.

[0039] After obtaining the predicted trajectory, the predicted area is determined according to this trajectory. The predicted area is usually a spatial range containing the predicted trajectory. The preset area range is set before the start of the unmanned aerial vehicle landing task, and it stipulates the target area that the unmanned aerial vehicle should reach during the landing process. By comparing the predicted area with the preset area range, the area deviation between them can be determined.

[0040] Furthermore, if the area deviation does not meet the preset conditions, for example, the magnitude of the area deviation exceeds the deviation threshold, it indicates that the predicted flight area has a large gap with the target area, and the unmanned aerial vehicle may not be able to land accurately according to the current state. The area deviation and the wind field parameters corresponding to the current layer of the unmanned aerial vehicle are input into the preset flight parameter adjustment model, and the model will output the flight adjustment parameters corresponding to the current layer of the unmanned aerial vehicle, and these parameters may include the speed adjustment amount, the direction adjustment angle, the attitude adjustment amplitude, etc.

[0041] Further, the flight parameters of the UAV are adjusted according to the obtained flight adjustment parameters. The flight control system of the UAV will change its own flight state according to these adjustment parameters, such as accelerating, decelerating, changing the flight direction, etc. Then, the adjusted flight parameters, altitude difference, and the wind field parameters corresponding to the current layer are input into the trajectory prediction model again, and the predicted trajectory of the UAV landing from the current position to the next layer is re-output. Through such continuous adjustment and prediction, the flight trajectory of the UAV gradually approaches the preset area range, and finally accurate landing is achieved.

[0042] In an implementation manner of the present application, a prediction area is obtained based on the predicted trajectory, specifically including: in the predicted trajectory, after a preset time interval, trajectory points are screened, and a trajectory point set is constructed based on the screened trajectory points. The trajectory point set is classified based on the spatial position. Based on the classified trajectory point set, the boundary contour of the predicted trajectory in space is constructed. Based on the highest point and the lowest point in the boundary contour, the boundary range in the vertical direction is determined, and, by using the convex hull algorithm, the smallest convex polygon containing all the trajectory points is determined, and the boundary of the smallest convex polygon is used as the boundary range of the predicted trajectory in the horizontal direction. Based on the boundary range in the vertical direction and the boundary range in the horizontal direction, the prediction range is obtained.

[0043] Specifically, trajectory points are screened every preset time interval to relatively evenly extract representative points from the continuous trajectory, and it is closely related to the time process of the UAV flight. These screened points constitute a trajectory point set. The trajectory point set constructed above is classified based on the spatial position, and the trajectory points are classified according to the horizontal direction and the vertical direction. The classification in the horizontal direction helps to analyze the movement path and range of the UAV on the plane, while the classification in the vertical direction allows us to understand the height change of the UAV.

[0044] Further, in the trajectory point set, the ordinate values of all points are found. The point corresponding to the maximum value is the highest point in the boundary contour, and the point corresponding to the minimum value is the lowest point. The ordinate values of these two points determine the boundary range in the vertical direction. The convex hull algorithm is used to determine the smallest convex polygon containing all the trajectory points, and the found smallest convex polygon is the boundary range of the predicted trajectory in the horizontal direction. Combining the boundary range in the vertical direction and the boundary range in the horizontal direction, the prediction range is obtained.

[0045] S103. The predicted trajectories corresponding to each landing layer are spliced to obtain the UAV landing trajectory.

[0046] In an implementation manner of the present application, the position deviation between the trajectory starting point of each landing layer and the center point of the target landing area is determined, and the direction deviation between adjacent predicted trajectories is determined. Based on the position deviation and the direction deviation, the predicted trajectories of each landing layer are translated and rotated in the spatial coordinate system so that the starting point and the ending point of the UAV at each landing layer are connected. Based on the connected predicted trajectories, Bezier curves are respectively constructed at each landing layer to generate the transition trajectories at each connection point through the Bezier curves. The adjusted predicted trajectories and the transition trajectories corresponding to each landing layer are spliced to obtain the UAV landing trajectory.

[0047] Specifically, for the trajectory starting points of each landing layer, it is necessary to determine the position deviation between them and the center point of the target landing area. The position deviation in the embodiments of the present application is obtained by calculating the coordinate difference between the two in the three-dimensional spatial coordinate system. This position deviation reflects the offset of the starting point of each landing layer relative to the target center point in space. Based on the change in direction when the predicted trajectory of the UAV transitions from the current landing layer to the predicted trajectory of the next landing layer, the direction deviation is determined. Using the position deviation, the predicted trajectory of each landing layer is translated in space, and based on the direction deviation, the predicted trajectory is rotated. Through the translation and rotation operations, it is finally realized that the starting point and the ending point of the UAV at each landing layer can be smoothly connected, making the flight trajectory of the UAV more coherent during the entire landing process.

[0048] Further, after the predicted trajectories of each landing layer are translated and rotated to achieve the connection of the starting point and the ending point, it is necessary to construct Bezier curves at each landing layer respectively. For example, the ending point of the predicted trajectory of the upper layer and the starting point of the predicted trajectory of the lower layer, as well as the auxiliary points at appropriate positions near these two points, are selected as the control points. By constructing the Bezier curve through these control points, a smooth transition trajectory can be generated between the two connection points. Such a transition trajectory can make the flight path of the UAV smoother when transitioning from one landing layer to another, avoiding sudden changes in direction or speed, and improving the flight stability and safety.

[0049] Further, starting from the first landing layer, the adjusted predicted trajectory and the corresponding transition trajectory are connected end to end. Then, by analogy, the adjusted predicted trajectories of the subsequent landing layers and the corresponding transition trajectories are spliced one by one in sequence. Through this splicing method, a complete, continuous and smooth UAV landing trajectory is finally obtained. This trajectory comprehensively considers the specific conditions of each landing layer and the transition between layers, and can guide the UAV to accurately and safely land in the target landing area.

[0050] S104. Obtain the wind field parameters corresponding to the target landing area, and determine the reference landing area based on the wind field parameters and the predicted trajectory.

[0051] In an implementation manner of the present application, feature extraction is performed on the predicted trajectory. Among them, the extracted features include at least one of the trajectory curvature change value, the speed change trend, and the spatial distribution of trajectory points at different time periods. Based on each time step of the predicted trajectory, according to the wind field parameters at the current position of the UAV, the position data of the UAV at the next moment is generated through Monte Carlo simulation. The generated position data at the next moment is classified by a vector machine. The position data at the next moment that falls within the safe range of the target landing area is marked as positive-class data, and other position data at the next moment is marked as negative-class data. Based on the positive-class data, a reference landing trajectory is determined.

[0052] Specifically, feature extraction is performed on the predicted trajectory. Among them, the trajectory curvature reflects the degree of trajectory bending. By calculating the curvature at different points on the predicted trajectory and analyzing the change of these curvatures with the order of trajectory points, the trajectory curvature change value is obtained; the speed data at different time steps on the predicted trajectory is obtained, for example, it is obtained whether the speed is gradually increasing, gradually decreasing or remaining relatively stable, so as to determine the speed change trend; the distribution of trajectory points in three-dimensional space at different time periods of the predicted trajectory is studied.

[0053] Furthermore, according to the wind field parameters at the current position of the UAV, including wind speed, wind direction, etc., the position data of the UAV at the next moment is generated by Monte Carlo simulation. Due to the uncertainty of the wind field, this uncertainty is considered in each simulation by randomly generating wind speed and wind direction change values that conform to the statistical characteristics of the wind field. Combining the current flight speed, direction and position information of the UAV, the possible position of the UAV at the next moment under this wind field condition is calculated through kinematic formulas.

[0054] Furthermore, a vector machine is used to classify the position data at the next moment generated by Monte Carlo simulation. Specifically, the safe range of the target landing area in the embodiments of the present application is a preset three-dimensional space area, which considers factors such as the space required for the safe landing of the UAV and the possible error range. Then, the position data at the next moment that falls within the safe range of the target landing area is marked as positive-class data, which means that these positions meet the requirements for safe landing; while other position data at the next moment that is not within this range is marked as negative-class data. The vector machine constructs a classification model by learning these marked data. This model can judge whether the input position data at the next moment belongs to the positive class or the negative class according to the characteristics of the data. For example, for a newly generated position data at the next moment, the vector machine calculates the relationship with the boundary of the safe range of the target landing area according to its position coordinates, and determines it as the positive class if it is within the range, otherwise determines it as the negative class.

[0055] Further, after classifying the position data of the next moment, a reference landing trajectory is determined based on the data marked as the positive class. The positive class data represents the position of the UAV at the next moment when it can safely land within the target area under various simulated wind field conditions. Connecting these positive class data in chronological order forms a continuous trajectory, which is the reference landing trajectory. The reference landing trajectory in the embodiments of the present application comprehensively considers the uncertainty of the wind field and the requirements for safe landing, providing a relatively reliable reference path for the actual landing of the UAV, enabling the UAV to have a greater probability of safely landing within the target area in a complex and changeable wind field environment.

[0056] S105. Determine the position deviation between the target landing area and the reference landing area, and based on the position deviation, determine the flight parameters to be adjusted for the UAV, so as to adjust the landing trajectory of the UAV based on the flight parameters to be adjusted for the UAV.

[0057] In an implementation manner of the present application, a three-dimensional space coordinate system is constructed based on the target landing area and the reference landing area. Based on the flight state of the UAV and environmental information, weight distribution is performed on the pseudorange measured by the UAV's GNSS and the differential correction data of RTK. The position deviation is determined in the three-dimensional space coordinate through the weighted data; wherein, the position deviation includes translational deviation and rotational deviation. The position deviation, the current pose of the UAV, and the wind field parameters corresponding to the target landing area are input into a preset target parameter adjustment model, so as to output the flight parameters to be adjusted for the UAV through the preset target parameter adjustment model.

[0058] Specifically, taking the center point of the landing platform as the origin, a coordinate system is constructed. Since the flight state of the UAV includes information such as flight altitude, speed, acceleration, and attitude angle. For example, when the flight altitude of the UAV is relatively high, the GNSS signal is relatively stable, and the pseudorange data measured by it may be highly reliable; while when the UAV approaches the target landing area, the flight altitude is low and the speed gradually decreases, the differential correction data of RTK is more critical for accurately determining the position. Therefore, considering the different requirements for positioning data in different stages of the flight state, these factors need to be comprehensively considered, and different weight values are assigned to the pseudorange measured by the UAV's GNSS and the differential correction data of RTK respectively. These weight values will be dynamically adjusted with the changes of the flight state and the environment to ensure that both types of data can be optimally utilized to determine the accurate position of the UAV in different situations.

[0059] Further, the GNSS measured pseudorange and RTK differential correction data after weight assignment are fused and calculated. According to the weight ratio, operations such as weighted summation are performed on the two types of data to obtain a result that comprehensively reflects the position of the UAV. The fused position coordinates are compared with the ideal position coordinates of the target landing area in the three-dimensional space coordinate system. Based on the comparison result, the obtained position deviation, the current UAV pose, and the wind field parameters corresponding to the target landing area are input into the preset target parameter adjustment model together. After the operation of the model, the flight parameters to be adjusted for the UAV are finally output. These parameters include the flight speed adjustment amount, the attitude angle adjustment amount, such as the adjustment values of the pitch angle, roll angle, and yaw angle, the acceleration adjustment amount, etc. The flight control system of the UAV can adjust the flight state of the UAV according to these flight parameters to be adjusted, so that it gradually approaches the target landing area and completes the landing with the correct attitude and speed.

[0060] In an implementation manner of the present application, based on the UAV flight state and environmental information, weight assignment is performed on the obtained GNSS measured pseudorange of the UAV and the RTK differential correction data, specifically including:

[0061] Through the function:

[0062]

[0063] Determine the weight of the GNSS measured pseudorange;

[0064] Based on the function:

[0065] W RTK = 1 - W GNSS ;

[0066] Determine the weight of the RTK differential correction data;

[0067] where W GNSS is the weight of the GNSS measured pseudorange; S GNSS is the GNSS signal strength; S RTK is the RTK signal strength; γ is the first adjustment coefficient; V is the flight speed; V max is the maximum flight speed of the UAV; δ is the second adjustment coefficient; H is the flight height; H max is the maximum flight height of the UAV; ε is the third adjustment coefficient; σ is the terrain undulation standard deviation; β is the fourth adjustment coefficient; α is the fifth adjustment coefficient; W RTK is the weight of the RTK differential correction data.

[0068] In an implementation manner of the present application, based on the flight parameters to be adjusted, the adjustment amount corresponding to the next trajectory point is determined, so as to re-plan the coordinates of the next trajectory point based on the adjustment amount. Based on each adjusted next trajectory point, trajectory discrete points are constructed, and the trajectory discrete points are connected into a curve through a spline curve fitting algorithm. Based on the connected curve, the adjusted UAV landing trajectory is obtained.

[0069] Specifically, according to the flight parameters to be adjusted and in combination with the current flight state of the UAV, such as the current position, speed, attitude, etc., the adjustment amount corresponding to the next trajectory point is determined. The adjustment amount in the embodiments of the present application refers to the specific adjustment value of the flight state required for the UAV to move from the current position to the next trajectory point. Based on the determined adjustment amount, the coordinates of the next trajectory point are re-planned. For example, according to the current speed, acceleration, and speed adjustment amount, the displacement at the next moment can be calculated, and then combined with the current position and attitude angle adjustment amount, the new coordinate position is determined. In this way, the coordinates of each next trajectory point are re-planned to meet the requirements of the flight parameters to be adjusted.

[0070] Further, after re-planning the coordinates of all the next trajectory points, these adjusted next trajectory points constitute trajectory discrete points. These discrete points represent the possible positions of the UAV at different times in three-dimensional space. Using the spline curve fitting algorithm, all the trajectory discrete points are connected to form a continuous and smooth curve. The connected curve is the adjusted UAV landing trajectory. By following this trajectory, the UAV can more effectively avoid obstacles, adapt to different wind field conditions, ensure accurate landing in the target landing area, and improve the success rate and safety of UAV landing.

[0071] S106. Control the UAV to land in the target landing area through the adjusted UAV landing trajectory.

[0072] In an implementation manner of the present application, according to the deviation analysis result between the current position of the UAV and the landing trajectory, the flight control system generates corresponding flight control instructions. The UAV adjusts its attitude and speed according to the flight control instructions. When the UAV approaches the target landing area, the flight control system makes a judgment according to the preset safety standards and landing conditions. For example, when the altitude of the UAV decreases to a certain extent, the speed decreases to an appropriate value, and the attitude is stable, the flight control system considers that the landing conditions are met and allows the UAV to contact the ground to complete the landing. If it is found that some parameters do not meet the requirements when approaching the landing area, the flight control system may continue to make adjustments, or issue an alarm and take corresponding emergency measures to ensure the safety of the UAV and the surrounding environment.

[0073] Figure 2This is a schematic structural diagram of a fixed-point landing control device for an unmanned aerial vehicle provided by an embodiment of the present application. As Figure 2 shown, the fixed-point landing control device 200 for an unmanned aerial vehicle includes: at least one processor 201; and a memory 202 communicatively connected to the at least one processor 201; wherein, the memory 202 stores instructions executable by the at least one processor 201, and when the instructions are executed by the at least one processor 201, the at least one processor 201 is capable of: when receiving a landing request of the unmanned aerial vehicle, obtaining the relative position between the unmanned aerial vehicle and the target landing area, dividing the landing space into multiple landing layers based on the relative position, and determining the height difference between the landing layers; constructing a predicted trajectory corresponding to each landing layer based on the height difference, the current flight parameters corresponding to the unmanned aerial vehicle, and the wind field parameters corresponding to each landing layer; splicing the predicted trajectories corresponding to each landing layer to obtain a landing trajectory of the unmanned aerial vehicle; obtaining the wind field parameters corresponding to the target landing area, and determining a reference landing area based on the wind field parameters and the predicted trajectory; determining the position deviation between the target landing area and the reference landing area, and determining the flight parameters to be adjusted for the unmanned aerial vehicle based on the position deviation, so as to adjust the landing trajectory of the unmanned aerial vehicle based on the flight parameters to be adjusted for the unmanned aerial vehicle; and controlling the unmanned aerial vehicle to land on the target landing area through the adjusted landing trajectory of the unmanned aerial vehicle.

[0074] A non-volatile computer storage medium provided by an embodiment of the present application stores computer-executable instructions, and the computer-executable instructions are configured to: when receiving a landing request of the unmanned aerial vehicle, obtain the relative position between the unmanned aerial vehicle and the target landing area, divide the landing space into multiple landing layers based on the relative position, and determine the height difference between the landing layers; construct a predicted trajectory corresponding to each landing layer based on the height difference, the current flight parameters corresponding to the unmanned aerial vehicle, and the wind field parameters corresponding to each landing layer; splice the predicted trajectories corresponding to each landing layer to obtain a landing trajectory of the unmanned aerial vehicle; obtain the wind field parameters corresponding to the target landing area, and determine a reference landing area based on the wind field parameters and the predicted trajectory; determine the position deviation between the target landing area and the reference landing area, and determine the flight parameters to be adjusted for the unmanned aerial vehicle based on the position deviation, so as to adjust the landing trajectory of the unmanned aerial vehicle based on the flight parameters to be adjusted for the unmanned aerial vehicle; and control the unmanned aerial vehicle to land on the target landing area through the adjusted landing trajectory of the unmanned aerial vehicle.

[0075] Each embodiment in the present application is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiments.

[0076] The foregoing are only examples of the present application and are not intended to limit the present application. For those skilled in the art, various modifications and variations can be made to the embodiments of the present application. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for controlling a drone's fixed-point landing, characterized in that: The method comprises: When receiving a landing request of a drone, obtaining a relative position between the drone and a target landing area, dividing the landing space into a plurality of landing layers based on the relative position, and determining a height difference between each landing layer; Based on the height difference, the current flight parameters corresponding to the UAV, and the wind field parameters corresponding to each landing layer, construct the predicted trajectory corresponding to each landing layer; The predicted trajectories corresponding to the landing layers are spliced ​​to obtain the landing trajectory of the UAV; Acquire wind field parameters corresponding to the target landing area, and determine a reference landing area based on the wind field parameters and the predicted trajectory; Determine a position deviation between the target landing area and the reference landing area, and determine a flight parameter of the drone to be adjusted based on the position deviation, so as to adjust the landing trajectory of the drone based on the flight parameter of the drone to be adjusted; The UAV is controlled to land in the target landing area through the adjusted landing trajectory of the UAV.

2. A method for controlling a drone's fixed-point landing according to claim 1, characterized in that: The step of constructing predicted trajectories corresponding to the landing layers based on the height difference, the current flight parameters corresponding to the UAV, and the wind field parameters corresponding to the landing layers specifically includes: Input the current flight parameters, the height difference, and the wind field parameters corresponding to the current layer of the drone into a preset trajectory prediction model, so as to output a predicted trajectory of the drone landing from the current position to the next layer through the trajectory prediction model; Obtaining a predicted area based on the predicted trajectory, and determining a regional deviation between the predicted area and a preset regional range; When the regional deviation does not meet the preset conditions, the regional deviation and the wind field parameter corresponding to the current layer of the UAV are input into a preset flight parameter adjustment model, so as to output the flight adjustment parameter corresponding to the current layer of the UAV through the preset flight parameter adjustment model; Based on the flight adjustment parameters corresponding to the current layer, the flight parameters of the UAV are adjusted, and the predicted trajectory of the UAV landing at the current position to the next layer is re-output through the adjusted flight parameters.

3. A method for controlling a drone to land at a fixed point according to claim 2, characterized in that: The obtaining of the predicted area based on the predicted trajectory specifically includes: In the predicted trajectory, after each preset time interval, trajectory points are screened, and a trajectory point set is constructed based on the screened trajectory points; Classifying the set of trajectory points based on spatial positions; Based on the classified trajectory point set, construct the boundary contour of the predicted trajectory in space; Determine the vertical boundary range based on the highest point and the lowest point in the boundary contour; And, determining the minimum convex polygon containing all the trajectory points by a convex hull algorithm, and taking the boundary of the minimum convex polygon as the boundary range of the predicted trajectory in the horizontal direction; The prediction range is obtained based on the boundary range in the vertical direction and the boundary range in the horizontal direction.

4. The method for controlling the fixed-point landing of an unmanned aerial vehicle according to claim 1, characterized in that: The predicted trajectories corresponding to the landing layers are spliced ​​to obtain the landing trajectory of the drone, specifically including: Determining the position deviation between the starting point of the trajectory of each landing layer and the center point of the target landing area, and determining the direction deviation between adjacent predicted trajectories; Based on the position deviation and the direction deviation, the predicted trajectory of each landing layer is translated and rotated in a spatial coordinate system so that the UAV is connected at the starting point and the ending point of each landing layer; Based on the predicted trajectory after connection, constructing Bezier curves in each landing layer respectively, so as to generate transition trajectories at each connection point through the Bezier curves; The adjusted predicted trajectories and transition trajectories corresponding to each landing layer are spliced ​​to obtain the landing trajectory of the UAV.

5. The method for controlling the fixed-point landing of an unmanned aerial vehicle according to claim 1, characterized in that: The obtaining of wind field parameters corresponding to the target landing area and determining a reference landing area based on the wind field parameters and the predicted trajectory specifically includes: Extracting features from the predicted trajectory; wherein the extracted features include at least one of a trajectory curvature change value, a speed change trend, and a spatial distribution of trajectory points in different time periods; Based on each time step of the predicted trajectory, the position data of the drone at the next moment is generated through Monte Carlo simulation according to the wind field parameters at the current position of the drone; Classifying the generated next moment position data by a vector machine, marking the next moment position data that falls within the safety range of the target landing area as positive data, and marking other next moment position data as negative data; Based on the positive data, the reference landing trajectory is determined.

6. The method for controlling the fixed-point landing of an unmanned aerial vehicle according to claim 1, characterized in that: The determining of the position deviation between the target landing area and the reference landing area, and determining the flight parameters of the UAV to be adjusted based on the position deviation, specifically includes: Constructing a three-dimensional space coordinate system based on the target landing area and the reference landing area; Based on the UAV flight status and environmental information, weights are allocated for the pseudo-ranges measured by the UAV GNSS and the differential correction data of RTK; Determine the position deviation in the three-dimensional space coordinates through the weighted data; wherein the position deviation includes a translation deviation and a rotation deviation; The position deviation, the current UAV posture and the wind field parameters corresponding to the target landing area are input into a preset target parameter adjustment model, so as to output the flight parameters to be adjusted of the UAV through the preset target parameter adjustment model.

7. The method for controlling the fixed-point landing of an unmanned aerial vehicle according to claim 1, characterized in that: The method of weighting the acquired pseudo-range measured by the UAV GNSS and the differential correction data of RTK based on the UAV flight status and environmental information specifically includes: Through the function: Determine the weight of the pseudorange measured by GNSS; Function-based: IN RTK =1-W GNSS ; Determine the weight of RTK differential correction data; Among them, W GNSS S is the weight of the GNSS pseudorange measurement; GNSS is the GNSS signal strength; S RTK is the RTK signal strength; γ is the first adjustment coefficient; V is the flight speed; V max is the maximum flight speed of the UAV; δ is the second adjustment coefficient; H is the flight altitude; H max is the maximum flight altitude of the UAV; ε is the third adjustment coefficient; σ is the standard deviation of terrain undulation; β is the fourth adjustment coefficient; α is the fifth adjustment coefficient; W RTK The weight of the RTK differential correction data.

8. The method for controlling the fixed-point landing of an unmanned aerial vehicle according to claim 1, characterized in that: The adjusting the landing trajectory of the drone based on the flight parameters to be adjusted of the drone specifically includes: Based on the flight parameter to be adjusted, determining the amount to be adjusted corresponding to the next trajectory point, so as to replan the coordinates of the next trajectory point based on the amount to be adjusted; Based on the adjusted next trajectory points, constructing trajectory discrete points, and connecting the trajectory discrete points into a curve through a spline curve fitting algorithm; Based on the connected curves, an adjusted landing trajectory of the UAV is obtained.

9. A UAV fixed-point landing control device, characterized in that: The device comprises a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device is triggered to execute the method according to any one of claims 1 to 8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions can execute the method according to any one of claims 1 to 8.

Images