Landing control method and system for vertical take-off and landing unmanned aerial vehicle
Through drone lidar scanning and obstacle point cloud density search algorithms, the drone landing path is generated and optimized, solving the problem of accurate landing of drones in complex environments and improving the real-time and accuracy of landing.
Patent Information
- Application Number
- CN202411129812.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing drone landing control methods are difficult to achieve accurate landing in complex environments such as dense forests, and the existing path planning technology has not been effectively integrated into global path planning, resulting in difficulty in ensuring real-time and optimal paths.
The onboard lidar scanning environment on the drone is adopted to build a local map through point cloud data, and a three-dimensional network is divided based on the obstacle point cloud density search algorithm to generate the initial polyline landing trajectory, and the final target polyline landing trajectory is output through trajectory optimization processing. Avoid obstacle collisions.
It realizes accurate landing of drones in complex environments, reduces workload and computing power consumption, avoids attitude imbalance caused by drones due to excessive dive turning angles, and improves the real-time and accuracy of landing.
Smart Images

Figure CN119002543A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of unmanned aerial vehicle flight control, and in particular to a landing control method for a vertical take-off and landing unmanned aerial vehicle. Background Art
[0002] When the drone completes its work and lands, existing path planning technologies usually do not integrate the movement of the drone into the global path planning, making it difficult to ensure real-time and optimal path during actual drone operation.
[0003] When planning a local path, it is necessary to track the global path while avoiding objects, and in the drone's map, determine the drone's position to complete takeoff and landing;
[0004] However, the local path planning is more demanding to quickly control the drone to achieve precise landing in a short time. As we all know, by understanding the terrain and obstacles in the landing area, it is necessary to use adaptive control technology to adjust the trajectory in real time, so that the drone can avoid obstacles in the landing trajectory when landing; however, due to the large number of obstacles and the uncertainty of obstacles, if a large number of obstacle point clouds are used to identify the obstacle type and then control the landing, the workload is obviously huge. For example, in a dense forest, the trees are dense, which poses a great challenge to the generation of the landing trajectory of the drone, and a more accurate and computationally efficient control adjustment of the landing trajectory is required. Summary of the invention
[0005] The object of the present invention is to provide a vertical take-off and landing UAV landing control method, comprising the following steps:
[0006] Use the airborne laser radar on the drone to scan environmental targets and collect point cloud data; build a local map of the current surrounding environment based on the point cloud data;
[0007] Get the starting position of the current drone and the position of the current helipad, use the starting position of the current drone as the starting point of the trajectory, and use the position of the current helipad as the end point of the trajectory;
[0008] Based on the density search algorithm of the obstacle point cloud, the landing path processing operation of the current UAV is generated by combining the trajectory starting point, trajectory end point and the local map of the current surrounding environment.
[0009] Preferably, the density search algorithm based on the obstacle point cloud is combined with the trajectory starting point, trajectory end point and the local map of the current surrounding environment to generate the landing path processing operation of the current drone, specifically including:
[0010] Dividing the local map of the current surrounding environment into a three-dimensional network;
[0011] According to the density of the obstacle point cloud, the three-dimensional space is divided into a complex obstacle area and an open obstacle area;
[0012] Preset the obstacle search step length; search for discrete nodes for nodes inside the complex obstacle area and the open obstacle area, connect the discrete nodes, and obtain an initial broken line landing trajectory;
[0013] Call the current initial broken line landing trajectory, perform trajectory optimization processing on the initial broken line landing trajectory, and output the final target broken line landing trajectory.
[0014] Preferably, searching for discrete nodes for nodes inside the obstacle complex area and the obstacle open area specifically includes the following operations:
[0015] In an obstacle-free area, a first preset step size is used to find nodes; and the ratio of the number of nodes found in the current obstacle-free area is set as a first target ratio, and then in an obstacle-complex area, a second preset step size is used to find nodes; and the ratio of the number of nodes found in the current obstacle-complex area is set as a second target ratio;
[0016] The first preset step length is set to be larger than the second preset step length, and the first target ratio is smaller than the second target ratio.
[0017] Preferably, the initial broken line landing trajectory is subjected to trajectory optimization processing to output the final target broken line landing trajectory. The specific operation steps are as follows:
[0018] Get the broken line formed by any two adjacent straight line segments in the current initial broken line landing trajectory, and then optimize the broken line of the current segment: determine the broken line endpoints at both ends of the current two adjacent straight line segments, determine the position coordinates of the three broken line endpoints formed at both ends of the adjacent straight line segments, and calculate the broken line turning angle through the position coordinates of the above three broken line endpoints; determine whether the broken line turning angle is greater than or equal to the first threshold value, if so, determine that the broken line of the previous segment is a non-optimized broken line, and then determine new nodes according to a preset distance in the direction of horizontally away from obstacles at the connection points between adjacent straight line segments, connect the new nodes with the end starting point and end point of the broken line of the current segment to form a new broken line, and retain the optimized broken line of the current segment; if not, if the current broken line turning angle is less than the first threshold value, retain the broken line of the current segment;
[0019] Traversing the broken lines formed by multiple adjacent straight line segments in the current broken line landing trajectory, and performing the above processing to obtain the broken line landing trajectory of the first screening;
[0020] The straight line segments in the first screened broken line landing trajectory are optimized for judging obstacles to obtain the broken line landing trajectory of the target.
[0021] Preferably, the optimization of judging obstacles on the straight line segments in the first screened broken line landing trajectory to obtain the broken line landing trajectory of the target specifically includes:
[0022] Mark the straight line segments of the first filtered broken line landing trajectory to obtain the first straight line segment, the second straight line segment, ... the Nth straight line segment, and so on;
[0023] First, the path points on the first straight line segment of the first screened broken line landing trajectory are obtained; in particular, the path points at the beginning and the end are obtained; the path points at both ends of the first straight line segment are used as the end starting point and the end point, and it is directly determined whether the first straight line segment crosses an obstacle. If so, the current straight line segment is discarded, and the path of the first straight line segment is replanned; if not, the current first straight line segment is retained, that is, the current straight line segment is verified to be a qualified straight line segment;
[0024] The first straight line segment to the Nth straight line segment are gradually traversed and verified until the broken line segment of the broken line landing trajectory currently screened for the first time is optimized to determine the obstacles, and the broken line landing trajectory of the final target is obtained.
[0025] Preferably, the replanning of the first straight line segment path specifically includes:
[0026] Execute the first priority level connection, generate the current line with the end starting point of the first straight line segment and the end point of the second straight line segment, directly determine whether the current line passes through an obstacle, if so, discard the current line, and execute step S3432 according to the second priority level growth path; if not, keep the current line as a qualified straight line segment;
[0027] Execute the second priority level connection, generate the current line from the end starting point of the first straight line segment to other path points between the second straight line segment, and directly determine whether the current line passes through an obstacle. If so, discard the current line; otherwise, retain the current line as a qualified straight line segment and regenerate the second straight line segment.
[0028] Preferably, when executing the second priority level connection, the current connection is generated from the end starting point of the first straight line segment to other path points between the second straight line segment; wherein, the position selection of other path points includes, starting from the end point of the second straight line segment as the starting point, searching for other path points at intervals of a preset step distance in the direction of the vector formed by the end point of the second straight line segment to the end starting point of the second straight line segment, ensuring that the line between the current other path point and the end starting point of the first straight line segment does not pass through an obstacle, and finally determining that the current other path point between the end starting point of the first straight line segment and the second straight line segment generates the current connection as the final first straight line segment.
[0029] Preferably, the end starting point and the end point on the straight line segment of the target broken line landing trajectory are obtained, and three control stages are determined for the current straight line segment, including an acceleration control stage, a uniform speed control stage, and a deceleration control stage:
[0030] First, determine the total length L of the current straight line segment; retrieve the control parameter set corresponding to the total length L of the current straight line segment within the standard range in the database;
[0031] The control parameter set includes the starting acceleration a0 of the acceleration control phase of the straight line segment, the deceleration a1 of the deceleration control phase, v max is the maximum speed, T is the total control time interval, the starting speed v0 at the end starting point, and the ending speed v1 at the end point; and T = t1 + t2 + t3;
[0032] Then call the control parameter set to execute the speed control process of the current straight line segment;
[0033] The calculation of the acceleration phase determines the time t1 of the acceleration phase.
[0034] Calculation of the uniform speed phase: The time of the uniform speed phase is t2, t2 = T-t1-t3;
[0035] Calculation of the deceleration phase The time of the deceleration phase t3,
[0036] Accordingly, the present invention also provides a vertical take-off and landing UAV landing control system, comprising: a collection module; a position module; a take-off and landing module;
[0037] The acquisition module is used to use the airborne laser radar on the drone to scan environmental targets and collect point cloud data; based on the point cloud data, a local map of the current surrounding environment is constructed;
[0038] The position module is used to obtain the starting position of the current drone and the position of the current apron, and use the starting position of the current drone as the starting point of the trajectory and the position of the current apron as the end point of the trajectory;
[0039] The take-off and landing module generates the landing path processing operation of the current UAV based on the density search algorithm of the obstacle point cloud, and combines the trajectory starting point, trajectory end point and the local map of the current surrounding environment.
[0040] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:
[0041] It can be seen from the analysis of the above-mentioned vertical take-off and landing UAV landing control method provided by the present invention that in specific applications, point cloud data is obtained by scanning through the laser radar on the UAV, the point cloud data is pre-processed, and a local map of the current surrounding environment is constructed, so that the scene of the surrounding environment of the UAV flight can be understood;
[0042] Furthermore, the local map of the current surrounding environment is divided into a three-dimensional network, and the three-dimensional network is divided into a complex obstacle area and an open obstacle area according to the density of the obstacle point cloud, so that the possible gathering places of obstacles can be understood; and by setting the obstacle search step, the complex obstacle area and the open obstacle area of the obstacle gathering place are searched for nodes, and the discrete nodes far away from the density of the obstacle point cloud are found, which can indicate that the discrete nodes are far away from the obstacles. The discrete nodes are connected to obtain the initial broken line landing trajectory;
[0043] Furthermore, the broken line formed by any two adjacent straight line segments in the current initial broken line landing trajectory is optimized, so that obstacles between the trajectories caused by omissions can be prevented when the discrete nodes in the above are screened. The route of the drone is screened again through optimization to prevent the drone from touching obstacles. First, the first threshold is judged by the broken line turning angle between the broken lines. If the angle of the broken line turning angle is greater than the first threshold, it means that the angle of the dive turning of the drone is too large when it descends, and it is easy to cause the attitude of the drone to be unbalanced. At this time, new nodes should be determined at a preset distance from the connecting points between adjacent straight line segments in the horizontal direction away from the obstacle to connect together to form a new broken line. In this way, the angle of the broken line turning angle between the broken lines is kept too small to avoid the attitude imbalance of the drone due to the excessive dive turning angle when it descends, and the broken line landing trajectory screened for the first time is obtained.
[0044] Furthermore, the straight line segments of the first screened broken line landing trajectory are marked, the end starting point A and the end point B on the first straight line segment are collected, and it is determined whether there is an obstacle on the line between the two points. If not, the first straight line segment is retained; if so, the first straight line segment is replanned; the first priority level connection is executed to connect the end starting point A and the end point C on the first straight line segment (that is, the end point C is the end point of the second straight line segment) to determine whether there is an obstacle. If so, a new first straight line segment is generated, which can not only reduce the landing flight distance of the drone, but also reduce the attitude deviation of the drone caused by the dive angle, and also reduce the descent time; if not, the second priority level connection is executed to collect the straight line between the second straight line segments and connect the node B' (A and B') to form a new first straight line segment, so that the first screened broken line landing trajectory can be optimized again starting from the first straight line segment until the final target broken line landing trajectory is obtained.
[0045] The vertical take-off and landing UAV landing control method adopted by the present invention does not perform point cloud recognition and classification on all obstacles, but uses a method of calculating obstacle point cloud density to screen the area, and then performs a step of planning the landing control path. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0047] Figure 1 A UAV landing flow chart provided for a vertical take-off and landing UAV landing control method of the present invention;
[0048] Figure 2 Trajectory generation flow chart provided for a vertical take-off and landing UAV landing control method of the present invention
[0049] Figure 3 A schematic diagram of an initial broken-line landing trajectory generated by a vertical take-off and landing UAV landing control method according to the present invention;
[0050] Figure 4 A flow chart for optimizing the initial broken line landing trajectory provided by a vertical take-off and landing UAV landing control method of the present invention;
[0051] Figure 5 A schematic diagram of a broken line with an excessively large turning angle provided by a vertical take-off and landing UAV landing control method of the present invention;
[0052] Figure 6 A schematic diagram of a broken line with a small turning angle provided by a vertical take-off and landing UAV landing control method of the present invention;
[0053] Figure 7 A flow chart of replanning the first straight line segment path provided for a vertical take-off and landing UAV landing control method of the present invention;
[0054] Figure 8 A schematic diagram of the connection between the replanned first straight line path A and B provided by a vertical take-off and landing UAV landing control method of the present invention;
[0055] Fig. 9 A schematic diagram of the connection between the re-planned first straight line path A and B' provided by a vertical take-off and landing UAV landing control method of the present invention;
[0056] Fig.10 A process diagram of a vertical take-off and landing UAV landing control system provided by the present invention;
[0057] Markings: acquisition module 10; location module 20; take-off and landing module 30. DETAILED DESCRIPTION
[0058] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.
[0059] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.
[0060] Embodiment 1
[0061] like Figure 1 As shown, the present invention proposes a vertical take-off and landing UAV landing control method, comprising the following operating steps:
[0062] S10: Use the airborne laser radar on the drone to scan environmental targets and collect point cloud data; construct a local map of the current surrounding environment based on the point cloud data;
[0063] It should be noted that the airborne laser radar on the drone scans environmental targets (i.e., objects in the surrounding environment) to obtain point cloud data of the current surrounding flight environment; the point cloud data of the two adjacent frames (i.e., the drone will continuously scan surrounding targets as time changes during flight) are preprocessed to obtain the preset object positioning feature points and point clouds of the two adjacent frames in the current surrounding environment; the preset object positioning feature points refer to the object feature points in a stable form selected from the point cloud data, such as feature points in the form of buildings; the preset object positioning feature points extracted from the point cloud data using the SLAM algorithm are used to construct a local map of the current surrounding environment. The use of the SLAM algorithm to construct a local map is a conventional technical means and will not be repeated here;
[0064] S20: Obtain the starting position of the current UAV and the position of the current apron, use the starting position of the current UAV as the starting point of the trajectory, and use the position of the current apron as the end point of the trajectory;
[0065] S30: Based on the density search algorithm of the obstacle point cloud, the landing path processing operation of the current UAV is generated in combination with the trajectory starting point, the trajectory end point and the called local map of the current surrounding environment.
[0066] It should be noted that the point cloud is obtained by preprocessing the point cloud data, and the obstacle point cloud is judged by the aggregation density of the point cloud (that is, because the obstacle has a large density and volume after the laser radar point cloud data is preprocessed to obtain the point cloud, the density of the generated point cloud will also be large and used as the obstacle point cloud). The density of the obstacle point cloud is used to search for the discrete point cloud between the trajectory starting point and the trajectory end point in the local map of the current surrounding environment, and connected together to form the landing path of the current drone;
[0067] like Figure 2 As shown, in the specific technical solution S30, the density search algorithm based on the obstacle point cloud is combined with the trajectory starting point, trajectory end point and the local map of the current surrounding environment to generate the landing path processing operation of the current drone, specifically including:
[0068] S31: Dividing the local map of the current surrounding environment into a three-dimensional network;
[0069] S32: Dividing the three-dimensional space into a complex obstacle area and an open obstacle area according to the density of the obstacle point cloud;
[0070] S33: Preset the obstacle search step length; search for discrete nodes (i.e., point cloud nodes that are not close to non-obstacles in discrete form) in the nodes inside the complex obstacle area and the open obstacle area (i.e., point cloud nodes of non-obstacles), connect the discrete nodes, and obtain an initial broken line landing trajectory;
[0071] S34: calling the current initial broken line landing trajectory, performing trajectory optimization processing operations on the initial broken line landing trajectory, and outputting the final target broken line landing trajectory.
[0072] It should be noted that there may be more obstacles in the complex obstacle area because the number of point clouds is too dense, while there may be fewer obstacles in the open obstacle area because the number of point clouds is small. The preset obstacle search step size is used to adjust a smaller obstacle search step size for the complex obstacle area. Because the number of point clouds in the complex obstacle area is too dense, a smaller obstacle search step size can be used to search more nodes and accurately determine discrete nodes. For the open obstacle area, a larger obstacle search step size is used because the number of point clouds in the open obstacle area is small, so the distribution of nodes is too loose, and discrete nodes can be searched more accurately, such as Figure 3 As shown; the discrete nodes indicate that many obstacles have been avoided, so the initial broken line landing trajectory is generated.
[0073] Through the above analysis and processing process, it can be achieved that an initial broken line landing trajectory is generated according to the obstacles in the local map of the current surrounding environment; in the path search process, the obstacles in the local map of the current surrounding environment are searched to determine the positions of the obstacles, thereby generating a landing trajectory; however, since there are obstacles in the local map of the current surrounding environment, in order to avoid the UAV from touching the obstacles when landing, an initial broken line landing trajectory is generated for the UAV by breaking line segments to avoid touching the obstacles, thereby achieving the purpose of landing the UAV.
[0074] When step S33 is executed, discrete nodes are searched for nodes inside the obstacle complex area and the obstacle open area, specifically including the following operations:
[0075] In a specific technical solution, discrete nodes are searched for nodes inside the complex obstacle area and the open obstacle area (i.e., point cloud nodes that are not obstacles). This technical solution also has further optimization processing methods.
[0076] S331: using a first preset step length to find nodes in an obstacle-free area; and setting the ratio of the number of nodes found in the current obstacle-free area to be a first target ratio, and then using a second preset step length to find nodes in an obstacle-complex area; and setting the ratio of the number of nodes found in the current obstacle-complex area to be a second target ratio;
[0077] The first preset step length is set to be larger than the second preset step length, and the first target ratio is smaller than the second target ratio.
[0078] It should be noted that the large step length node search operation mode should be adopted in the open area with obstacles and fewer nodes should be found in the open area, so that the path formed by the connection between the nodes (i.e., the point cloud nodes without obstacles) is more direct and the distance is shorter; on the contrary, the small step length node search should be adopted in the area with complex obstacles and more nodes should be found in the area with complex obstacles, so as to ensure a higher success rate of the nodes found by the search;
[0079] In summary, the space is divided into a three-dimensional grid, and each grid point represents a possible position. First, the three-dimensional grid is divided into an obstacle complex area with a large obstacle density and an obstacle open area with a small obstacle density; first, the search step size is adjusted according to the obstacle density, and a large step size is used to find nodes in the obstacle open area and find fewer nodes in the open area, and then a small step size is used to find nodes in the obstacle complex area to find more nodes. Then the selected target nodes are selected as discrete nodes; multiple discrete nodes are connected to each other to form multiple obstacle trajectories to be finally determined (i.e., the initial broken line landing trajectory) and integrated into the search space.
[0080] like Figure 4 As shown, specifically, in step S34, the initial broken line landing trajectory is subjected to trajectory optimization processing operation to output the final target broken line landing trajectory. The specific operation steps are as follows:
[0081] When step 34 is executed, the initial broken line landing trajectory is optimized and the final target broken line landing trajectory is output; thereby achieving a UAV landing trajectory that is more in line with the actual environment; it should be noted that after the UAV landing trajectory is generated, because the landing trajectory is a broken line segment, the landing trajectory of the broken line segment may be in the landing process, because it is necessary not only to avoid obstacles but also to pay attention to the landing dive angle should not be too large; therefore, it is necessary to perform secondary optimization processing on the landing trajectory, so that the landing trajectory can better conform to the UAV landing path in the UAV flight posture; the specific operation steps are as follows:
[0082] S341: Obtain the broken line formed by any two adjacent straight line segments in the current initial broken line landing trajectory, and then optimize the broken line of the current segment: determine the broken line endpoints at both ends of the current two adjacent straight line segments, determine the position coordinates of the three broken line endpoints formed at both ends of the adjacent straight line segments, and calculate the broken line turning angle through the position coordinates of the above three broken line endpoints; determine the broken line turning angle (that is, the angle of the drone when diving and turning should be limited and should not be too large; too large an angle is likely to cause posture imbalance, such as Figure 5 ) is greater than or equal to the first threshold. If so, the polyline of the previous segment is determined to be a non-optimized polyline. Then, the connecting points between adjacent straight line segments are connected horizontally away from the obstacle in accordance with the preset distance to determine a new node (i.e., a point cloud node that is not an obstacle). The new node is connected to the end starting point and the end point of the polyline of the current segment (i.e., as shown in FIG. 1 ). Figure 6 The end starting point A and the end point C of the polyline shown in the figure, where B and B' are both connecting points between adjacent straight line segments) form a new polyline, and retain the optimized polyline of the current segment; if not, if the current polyline angle is less than the first threshold, retain the polyline of the current segment;
[0083] Traversing the broken lines formed by multiple adjacent straight line segments in the current broken line landing trajectory, and performing the above processing to obtain the broken line landing trajectory of the first screening;
[0084] It should be noted that the broken line landing trajectory is optimized and screened through the coordinate positions of the broken line endpoints and the dive angles of the straight line segments of the broken line, and the broken line landing trajectory of the first screening is preliminarily obtained; the broken line landing trajectory of the first screening can be used to optimize the route of the current broken line landing trajectory, and generate a route that meets the broken line landing trajectory that the drone can complete autonomously; however, the study found that the initial broken line landing trajectory is mainly generated by the density calculation algorithm of the non-obstacle point cloud nodes, and there may still be obstacles on its trajectory, or the broken line landing trajectory may still pass through some obstacles. Therefore, it is necessary to use an optimization algorithm to verify the rationality of the initial broken line landing trajectory, see steps S342-S344 for details, and the straight line segments in the current first screened broken line landing trajectory are optimized to determine the obstacles, and the target broken line landing trajectory is obtained;
[0085] S342: marking straight line segments of the first screened broken line landing trajectory to obtain a first straight line segment, a second straight line segment, ... an Nth straight line segment, and so on;
[0086] S343: Firstly, the path points on the first straight line segment of the first screened broken line landing trajectory are obtained; in particular, the path points at the beginning and the end (the path points at the beginning and the end) are obtained; the path points at both ends of the first straight line segment are used as the end starting point (or root node) and the end point (that is, the end point of the first straight line segment is the end starting point of the second straight line segment), and it is directly determined whether the first straight line segment crosses the obstacle (that is, it is determined by determining whether the position coordinates on the first straight line segment are repeated with the obstacle coordinates). If so, the current straight line segment (the line between AB) is discarded, and the path of the first straight line segment is replanned; if not, the current first straight line segment is retained, that is, the current straight line segment is verified to be a qualified straight line segment;
[0087] It should be noted that the first screened broken line landing trajectory is composed of broken line segments connected together, so the first screened broken line landing trajectory is marked with straight line segments as: the first straight line segment, the second straight line segment, etc.;
[0088] S344: gradually traverse and verify the first straight line segment to the Nth straight line segment until the broken line segment of the broken line landing trajectory currently screened for the first time is optimized to determine obstacles, and the final target broken line landing trajectory is obtained.
[0089] Repeat the above steps S342-S344 to iteratively determine whether there are obstacles for all the broken line segments marked in the broken line landing trajectory of the first screening, plan a new broken line landing trajectory of the first screening, and generate the final optimized target broken line landing trajectory obtained by verification from the root node.
[0090] like Figure 7 As shown, in the above specific technical solution S343, the replanning of the first straight line segment path specifically includes:
[0091] S3431: Execute the first priority level connection, the end starting point of the first straight line segment and the end point of the second straight line segment generate the current line (that is, during the connection process, if an obstacle appears, it proves that the straight line segment is unqualified), directly determine whether the current line passes through the obstacle (that is, specifically determine whether the position coordinates on the current line are repeated with the obstacle coordinates). If so, the current line is discarded, and step S3432 is executed according to the second priority level growth path; otherwise, the current line is retained as a qualified straight line segment, such as Figure 8 As shown;
[0092] S3432: Execute the second priority level connection, generate the current line from the end starting point of the first straight line segment (i.e. A) to the other path points between the second straight line segment (i.e. BC straight line segment), directly determine whether the current line passes through the obstacle (i.e., specifically determine whether the position coordinates on the current line are repeated with the obstacle coordinates), if so, discard the current line; otherwise, retain the current line as a qualified straight line segment (i.e., the newly generated qualified first straight line segment, such as Fig. 9 As shown), and regenerate the second straight line segment (B'C is used as the regenerated second straight line segment);
[0093] It should be noted that when replanning the first straight line segment path, the first priority level connects AC; the second priority level connects other path points (or other nodes) between the straight line segments A to BC;
[0094] Verify and replan the first straight line segment and the Nth straight line segment of the first screened polyline landing trajectory to prevent the drone from hitting obstacles during landing;
[0095] Gradually traverse and verify the first straight line segment to the Nth straight line segment until the broken line segment of the broken line landing trajectory currently screened for the first time is optimized to determine the obstacles, and obtain the broken line landing trajectory of the final target;
[0096] If obstacles are found between the first straight line segments during the detection process, a new first straight line segment needs to be planned. The end starting point (i.e., the root node) of the first straight line segment is directly connected with the end point of the second straight line segment to generate a first preferred level first straight line segment. If obstacles also appear between the first preferred level first straight line segments, a second preferred level first straight line segment is generated. This can shorten the flight path of the drone, reduce the landing time of the drone, and achieve rapid landing. This can also shorten the flight time and achieve the purpose of rapid landing.
[0097] Preferably, in the process of executing the second priority level connection, other path points between the end starting point of the first straight line segment and the second straight line segment generate the current connection execution;
[0098] Among them, the position selection of other path points includes, starting from the end point of the second straight line segment as the starting point, searching for other path points in the vector direction formed by the end point of the second straight line segment to the end starting point of the second straight line segment by a preset step distance, ensuring that the line between the current other path point and the end starting point of the first straight line segment does not pass through an obstacle, and finally determining the current other path point between the end starting point of the first straight line segment and the second straight line segment to generate the current line as the final first straight line segment.
[0099] See also Fig. 9 In the specific technical solution, the second priority level connection is executed, and the current connection line is generated from the end starting point of the first straight line segment to other path points between the second straight line segment; among them, the other path points are preferably selected from the path points that are relatively close to the end point C, that is, starting from the end point C of the second straight line segment as the starting point, searching B' at an interval preset step distance according to the vector direction of CB', ensuring that the connection between point B' and the end starting point A does not pass through obstacles, and finally determining the path point B' between the end starting point A of the first straight line segment and the second straight line segment to generate the current connection as the final and closest first straight line segment.
[0100] The present invention proposes a method for controlling the landing of a vertical take-off and landing unmanned aerial vehicle to implement the target broken line landing trajectory of the calculated output, which is a simple and practical broken line landing control method particularly suitable for vertical take-off and landing unmanned aerial vehicles. Its target broken line landing trajectory is essentially composed of a large number of straight line segments, and the total length L of each straight line segment is different. The operation control of this straight line gradually descends in a gradient manner is relatively simple, and the flight trajectory smoothed by the traditional Bezier curve and B-Spline Curve algorithm is easier to control and the operation is simpler; the Bezier curve is a curve defined by control points, and the shape of the curve can be changed by adjusting the position of the control point. In the path planning of the unmanned aerial vehicle, each turning point of the broken line segment can be used as a control point, and a smooth curve can be generated using the Bezier curve algorithm. However, in order to further shorten the landing time and realize precise landing control, the control on the straight line segment path is planned as follows according to time and speed control. Obviously, this method of directly calling the corresponding control parameter set will reduce the difficulty of the landing trajectory control of the unmanned aerial vehicle, as shown in S401-S403.
[0101] After obtaining the final optimized target broken line landing trajectory, it also includes:
[0102] S40, performing a flight control processing operation on each straight line segment of the target broken line landing trajectory; that is, after obtaining the final optimized target broken line landing trajectory, the process further includes performing a flight control processing operation on each straight line segment of the target broken line landing trajectory:
[0103] S401, obtaining the end starting point and the end point on the straight line segment of the target broken line landing trajectory, and determining three control stages for the current straight line segment, including an acceleration control stage, a uniform speed control stage, and a deceleration control stage:
[0104] S402, where v0 and v1 are set to be the starting speed at the end point and the ending speed at the end point, a0 and a1 are the starting acceleration of the acceleration control phase and the ending acceleration of the deceleration control phase of the straight line segment, and v max is the maximum speed, T is the total control time interval, T = t1 + t2 + t3; L is the total length of the current straight line segment;
[0105] S403, first determine the total length L of the current straight segment; retrieve the control parameter set corresponding to the total length L of the current straight segment in the standard range in the database; that is, the total length L of each straight segment is different, but the database has preset the corresponding control parameter set when the total length L is 1-3 meters; the corresponding control parameter set when the total length L is 3-10 meters, etc.; therefore, the corresponding control parameter set can be determined according to the total length L of the current straight segment; then call the control parameter set to execute the speed control process of the current straight segment;
[0106] The control parameter set includes the starting acceleration a0 of the acceleration control phase of the straight line segment, the deceleration a1 of the deceleration control phase, v max is the maximum speed, T is the total control time interval (and T = t1 + t2 + t3), the starting speed v0 at the end starting point, and the ending speed v1 at the end point;
[0107] Calculation of the acceleration phase Determine the time t1 of the acceleration phase: Use the formula v = v0 + at, where a is the acceleration a0,
[0108] Calculation of the uniform speed phase The time t2 of the uniform speed phase is calculated by subtracting the time of the acceleration and deceleration phases from the total time T: t2 = T-t1-t3;
[0109] Calculation of the deceleration phase The time t3 of the deceleration phase is calculated using the deceleration a1 as follows:
[0110]
[0111] Calculation of total length L: The total length L can be calculated by adding the displacements of the three stages.
[0112] The displacement s can be calculated using the formula To calculate:
[0113] Acceleration phase displacement
[0114] Displacement in the uniform speed stage s2 = v max t2;
[0115] Displacement during deceleration phase
[0116] The total length L is equal to s1+s2+s3.
[0117] Embodiment 2
[0118] like Fig.10 As shown, the present invention also provides a vertical take-off and landing UAV landing control system, including: a collection module 10; a position module 20; a take-off and landing module 30;
[0119] The acquisition module 10 is used to use the airborne laser radar on the drone to scan the environmental target and acquire point cloud data; and to construct a local map of the current surrounding environment based on the point cloud data;
[0120] The position module 20 is used to obtain the starting position of the current drone and the position of the current apron, and use the starting position of the current drone as the starting point of the trajectory and the position of the current apron as the end point of the trajectory;
[0121] The take-off and landing module 30 generates a landing path processing operation for the current UAV based on a density search algorithm of the obstacle point cloud and in combination with the trajectory starting point, trajectory end point and the called local map of the current surrounding environment.
[0122] In summary, in a vertical take-off and landing UAV landing control method proposed in an example of the present invention, a laser radar on the UAV scans to obtain point cloud data, pre-processes the point cloud data, and constructs a local map of the current surrounding environment, so as to understand the scene of the surrounding environment of the UAV flight;
[0123] Furthermore, the local map of the current surrounding environment is divided into a three-dimensional network, and the three-dimensional network is divided into a complex obstacle area and an open obstacle area according to the density of the obstacle point cloud, so that the possible gathering places of obstacles can be understood; and by setting the obstacle search step, the complex obstacle area and the open obstacle area of the obstacle gathering place are searched for nodes, and the discrete nodes far away from the density of the obstacle point cloud are found, which can indicate that the discrete nodes are far away from the obstacles. The discrete nodes are connected to obtain the initial broken line landing trajectory;
[0124] Furthermore, the broken line formed by any two adjacent straight line segments in the current initial broken line landing trajectory is optimized, so that obstacles between the trajectories caused by omissions can be prevented when the discrete nodes in the above are screened. The route of the drone is screened again through optimization to prevent the drone from touching obstacles. First, the first threshold is judged by the broken line turning angle between the broken lines. If the angle of the broken line turning angle is greater than the first threshold, it means that the angle of the dive turning of the drone is too large when it descends, and it is easy to cause the attitude of the drone to be unbalanced. At this time, new nodes should be determined at a preset distance from the connecting points between adjacent straight line segments in the horizontal direction away from the obstacle to connect together to form a new broken line. In this way, the angle of the broken line turning angle between the broken lines is kept too small to avoid the attitude imbalance of the drone due to the excessive dive turning angle when it descends, and the broken line landing trajectory screened for the first time is obtained.
[0125] Furthermore, the straight line segments of the first screened broken line landing trajectory are marked, the end starting point A and the end point B on the first straight line segment are collected, and it is determined whether there is an obstacle on the line between the two points. If not, the first straight line segment is retained; if so, the first straight line segment is replanned; the first priority level connection is executed to connect the end starting point A and the end point C on the first straight line segment (that is, the end point C is the end point of the second straight line segment) to determine whether there is an obstacle. If so, a new first straight line segment is generated, which can not only reduce the landing flight distance of the drone, but also reduce the attitude deviation of the drone caused by the dive angle, and also reduce the descent time; if not, the second priority level connection is executed to collect the straight line between the second straight line segments and connect the node B' (A and B') to form a new first straight line segment, so that the first screened broken line landing trajectory can be optimized again starting from the first straight line segment until the final target broken line landing trajectory is obtained.
[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. A person skilled in the art may modify the technical solutions described in the above embodiments, or replace part or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A vertical take-off and landing UAV landing control method, characterized in that: The steps include: Use the airborne laser radar on the drone to scan environmental targets and collect point cloud data; build a local map of the current surrounding environment based on the point cloud data; Get the starting position of the current drone and the position of the current helipad, use the starting position of the current drone as the starting point of the trajectory, and use the position of the current helipad as the end point of the trajectory; Based on the density search algorithm of the obstacle point cloud, the landing path processing operation of the current UAV is generated by combining the trajectory starting point, trajectory end point and the local map of the current surrounding environment.
2. A vertical take-off and landing UAV landing control method according to claim 1, characterized in that: The density search algorithm based on the obstacle point cloud is combined with the trajectory starting point, trajectory end point and the local map of the current surrounding environment to generate the landing path processing operation of the current drone, specifically including: Dividing the local map of the current surrounding environment into a three-dimensional network; According to the density of the obstacle point cloud, the three-dimensional space is divided into a complex obstacle area and an open obstacle area; Preset the obstacle search step length; search for discrete nodes for nodes inside the complex obstacle area and the open obstacle area, connect the discrete nodes, and obtain an initial broken line landing trajectory; Call the current initial broken line landing trajectory, perform trajectory optimization processing on the initial broken line landing trajectory, and output the final target broken line landing trajectory.
3. A vertical take-off and landing UAV landing control method according to claim 2, characterized in that: Searching for discrete nodes for nodes in the obstacle complex area and the obstacle open area specifically includes the following operations: In an obstacle-free area, a first preset step length is used to find nodes; and the ratio of the number of nodes found in the current obstacle-free area is set as a first target ratio, and then in an obstacle-complex area, a second preset step length is used to find nodes; and the ratio of the number of nodes found in the current obstacle-complex area is set as a second target ratio; The first preset step length is set to be larger than the second preset step length, and the first target ratio is smaller than the second target ratio.
4. A vertical take-off and landing UAV landing control method according to claim 3, characterized in that: Perform trajectory optimization processing on the initial broken line landing trajectory and output the final target broken line landing trajectory. The specific steps are as follows: Get the broken line formed by any two adjacent straight line segments in the current initial broken line landing trajectory, and then optimize the broken line of the current segment: determine the broken line endpoints at both ends of the current two adjacent straight line segments, determine the position coordinates of the three broken line endpoints formed at both ends of the adjacent straight line segments, and calculate the broken line turning angle through the position coordinates of the above three broken line endpoints; determine whether the broken line turning angle is greater than or equal to the first threshold value, if so, determine that the broken line of the previous segment is a non-optimized broken line, and then determine new nodes according to a preset distance in the direction of horizontally away from obstacles at the connection points between adjacent straight line segments, connect the new nodes with the end starting point and end point of the broken line of the current segment to form a new broken line, and retain the optimized broken line of the current segment; if not, if the current broken line turning angle is less than the first threshold value, retain the broken line of the current segment; Traversing the broken lines formed by multiple adjacent straight line segments in the current broken line landing trajectory, and performing the above processing to obtain the broken line landing trajectory of the first screening; The straight line segments in the first screened broken line landing trajectory are optimized for judging obstacles to obtain the broken line landing trajectory of the target.
5. A vertical take-off and landing UAV landing control method according to claim 4, characterized in that: Optimizing the obstacle determination of the straight line segments in the first screened broken line landing trajectory to obtain the broken line landing trajectory of the target specifically includes: Mark the straight line segments of the first filtered broken line landing trajectory to obtain the first straight line segment, the second straight line segment, ... the Nth straight line segment, and so on; First, the path points on the first straight line segment of the first screened broken line landing trajectory are obtained; in particular, the path points at the beginning and the end are obtained; the path points at both ends of the first straight line segment are used as the end starting point and the end point, and it is directly determined whether the first straight line segment crosses an obstacle. If so, the current straight line segment is discarded, and the path of the first straight line segment is replanned; if not, the current first straight line segment is retained, that is, the current straight line segment is verified to be a qualified straight line segment; The first straight line segment to the Nth straight line segment are gradually traversed and verified until the broken line segment of the broken line landing trajectory currently screened for the first time is optimized to determine the obstacles, and the broken line landing trajectory of the final target is obtained.
6. A vertical take-off and landing UAV landing control method according to claim 5, characterized in that: The replanning of the first straight line segment path specifically includes: Execute the first priority level connection, generate the current line with the end starting point of the first straight line segment and the end point of the second straight line segment, directly determine whether the current line passes through an obstacle, if so, discard the current line, and execute step S3432 according to the second priority level growth path; if not, keep the current line as a qualified straight line segment; Execute the second priority level connection, generate the current line from the end starting point of the first straight line segment to other path points between the second straight line segment, and directly determine whether the current line passes through an obstacle. If so, discard the current line; otherwise, retain the current line as a qualified straight line segment and regenerate the second straight line segment.
7. A vertical take-off and landing UAV landing control method according to claim 6, characterized in that: During the execution of the second priority level connection, the current line is generated from the end starting point of the first straight line segment to other path points between the second straight line segment; wherein, the position selection of other path points includes, starting from the end point of the second straight line segment as the starting point, searching for other path points at intervals of a preset step distance in the direction of the vector formed by the end point of the second straight line segment to the end starting point of the second straight line segment, ensuring that the line between the current other path point and the end starting point of the first straight line segment does not pass through an obstacle, and finally determining that the current other path point between the end starting point of the first straight line segment and the second straight line segment generates the current line as the final first straight line segment.
8. A vertical take-off and landing UAV landing control method according to claim 7, characterized in that: After obtaining the final optimized target broken line landing trajectory, the flight control processing operation of each straight line segment of the target broken line landing trajectory is also included: Get the end starting point and end point on the straight line segment of the target broken line landing trajectory, and determine three control stages for the current straight line segment, including acceleration control stage, uniform speed control stage, and deceleration control stage: First, determine the total length L of the current straight line segment; retrieve the control parameter set corresponding to the total length L of the current straight line segment within the standard range in the database; The control parameter set includes the starting acceleration a0 of the acceleration control phase of the straight line segment, the deceleration a1 of the deceleration control phase, v max is the maximum speed, T is the total control time interval, the starting speed v0 at the end starting point, and the ending speed v1 at the end point; and T = t1 + t2 + t3; Then call the control parameter set to execute the speed control process of the current straight line segment; The calculation of the acceleration phase determines the time t1 of the acceleration phase. Calculation of the uniform speed phase: The time of the uniform speed phase is t2, t2 = T-t1-t3; Calculation of the deceleration phase The time of the deceleration phase t3, 9. A vertical take-off and landing UAV landing control system, characterized in that: include: Acquisition module; Location module; Take-off and landing module; The acquisition module is used to use the airborne laser radar on the drone to scan environmental targets and collect point cloud data; Build a local map of the current surrounding environment based on point cloud data; The position module is used to obtain the starting position of the current drone and the position of the current apron, and use the starting position of the current drone as the starting point of the trajectory and the position of the current apron as the end point of the trajectory; The take-off and landing module generates the landing path processing operation of the current UAV based on the density search algorithm of the obstacle point cloud, and combines the trajectory starting point, trajectory end point and the local map of the current surrounding environment.
Citation Information
Patent Citations
UAV path optimization method based on improved bidirectional rapidly-exploring random-tree algorithm
CN108681787A
Computer-assisted ascertainment of a movement of an apparatus
CN110501970A
Unmanned aerial vehicle non-cooperative target tracking system based on binocular vision
CN113467500A
Unmanned aerial vehicle system with intelligent obstacle avoidance and fixed-height cruising functions and control method
CN114859962A
Unmanned aerial vehicle optimal path selection method, terminal and storage medium
CN115047905A
Cited By
Unmanned aerial vehicle autonomous landing path planning method and system based on time-energy
CN119847188A
Time-Energy Based UAV Autonomous Landing Path Planning Method and System
CN119847188B
