Unmanned aerial vehicle recovery path adaptive planning method and system based on air-ground cooperation

Abstract

The application provides a UAV recovery path adaptive planning method and system based on air-ground cooperation, relates to the technical field of robot motion planning, and comprises the following steps: constructing a three-dimensional grid map of a task environment, acquiring the positions of a UAV and an unmanned vehicle and the maximum flight distance of the UAV; determining the responsibility distribution of the UAV and the unmanned vehicle in a recovery task based on the positions of the UAV and the unmanned vehicle and the maximum flight distance of the UAV, and obtaining an optimal UAV recovery scheme; selecting a corresponding path planning mode according to the optimal UAV recovery scheme, and determining the starting point, the ending point and the dimension of the search space of path planning based on the positions of the UAV and the unmanned vehicle; executing a path planning algorithm according to the selected path planning mode, and generating a final recovery path; and through air-ground cooperative decision and cross-dimension bidirectional joint path planning algorithm, the application realizes safe and efficient recovery of the UAV under the constraint of electric quantity, and significantly improves the task success rate and the system cooperation efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot motion planning, in particular to a method and system for adaptive planning of a UAV recovery path based on air-ground cooperation. BACKGROUND

[0002] A UAV is favored by many researchers and institutions due to its wide field of view and strong maneuverability. However, the limited endurance makes it difficult for the UAV to perform long-range and high-load tasks, which severely limits the application range of the UAV. An air-ground cooperation system can complete complex tasks beyond the individual capabilities of the UAV / car by combining the advantages of the UAV with the long endurance and strong load capacity of the unmanned vehicle. It has strong application potential in a wide range of scenarios such as field search and rescue, express delivery, and environmental detection. During task execution, the unmanned vehicle, as a task execution individual, can also serve as a task support platform for the air-ground cooperation system, providing a charging platform for the UAV, effectively expanding the task coverage range, and comprehensively improving the overall task flexibility of the system. This requires the UAV and the unmanned vehicle to interact and cooperate multiple times, and the UAV recovery is a key link to ensure the successful execution of the task.

[0003] In practical applications, the UAV and the unmanned vehicle usually undertake different sub-tasks in the air-ground cooperation task scenario, such as performing different regional, different type, or different stage tasks, to fully leverage the advantages of UAV and unmanned vehicle cooperation, decompose complex tasks, and improve the overall efficiency and quality of task execution. Therefore, when the UAV's power is insufficient, due to distance and environmental restrictions, the UAV may not be able to land on the unmanned vehicle before the power is depleted. How to achieve safe and efficient recovery under power constraints is a challenging problem in the task execution process of the air-ground cooperation system.

[0004] Existing UAV recovery methods are mostly single-platform perspective, separately planning the UAV or unmanned vehicle path and trajectory, lacking air-ground cooperation decision-making, resulting in insufficient depth of UAV and unmanned vehicle cooperation, and suboptimal recovery path in terms of cooperation, safety, or energy utilization. In addition, existing UAV recovery methods usually focus on the performance of landing platform visual recognition or landing control, and lack consideration of UAV power constraints, making it difficult to guarantee the success rate of the recovery task when the UAV's power is insufficient, and even causing serious accidents such as crashes due to power depletion. SUMMARY

[0005] To solve the above problems, the present application proposes a method and system for adaptive planning of a UAV recovery path based on air-ground cooperation, which realizes safe and efficient recovery of the UAV under power constraints through air-ground cooperation decision-making and cross-dimension bidirectional joint path planning algorithm, significantly improving the task success rate and system cooperation efficiency.

[0006] According to some embodiments, the present application adopts the technical solutions as follows:

[0007] The unmanned aerial vehicle recovery path adaptive planning method based on air-ground cooperation comprises:

[0008] A three-dimensional grid map of a task environment is constructed, and the positions of the unmanned aerial vehicle and the unmanned vehicle and the maximum flight distance of the unmanned aerial vehicle are obtained;

[0009] Based on the positions of the unmanned aerial vehicle and the unmanned vehicle and the maximum flight distance of the unmanned aerial vehicle, the responsibility distribution of the unmanned aerial vehicle and the unmanned vehicle in the recovery task is determined, and the best unmanned aerial vehicle recovery scheme is obtained;

[0010] According to the best unmanned aerial vehicle recovery scheme, a corresponding path planning mode is selected, and the starting point, the ending point and the dimension of the search space of the path planning are determined based on the positions of the unmanned aerial vehicle and the unmanned vehicle;

[0011] According to the selected path planning mode, a path planning algorithm is executed to generate the final recovery path;

[0012] Among them, when the responsibility distribution is that the unmanned aerial vehicle and the unmanned vehicle respectively undertake part of the recovery responsibility in the recovery scheme, a cross-dimension bidirectional joint path planning algorithm is adopted to generate a safe travel path of the unmanned aerial vehicle and the unmanned vehicle under the path distance constraint.

[0013] According to some embodiments, the present application adopts the technical solutions as follows:

[0014] The unmanned aerial vehicle recovery path adaptive planning system based on air-ground cooperation comprises:

[0015] The position acquisition module is configured to construct a three-dimensional grid map of a task environment, and obtain the positions of the unmanned aerial vehicle and the unmanned vehicle and the maximum flight distance of the unmanned aerial vehicle;

[0016] The responsibility distribution module is configured to determine the responsibility distribution of the unmanned aerial vehicle and the unmanned vehicle in the recovery task based on the positions of the unmanned aerial vehicle and the unmanned vehicle and the maximum flight distance of the unmanned aerial vehicle, and obtain the best unmanned aerial vehicle recovery scheme;

[0017] The mode selection module is configured to select a corresponding path planning mode according to the best unmanned aerial vehicle recovery scheme, and determine the starting point, the ending point and the dimension of the search space of the path planning based on the positions of the unmanned aerial vehicle and the unmanned vehicle;

[0018] The path planning module is configured to execute a path planning algorithm according to the selected path planning mode to generate the final recovery path;

[0019] Among them, when the responsibility distribution is that the unmanned aerial vehicle and the unmanned vehicle respectively undertake part of the recovery responsibility in the recovery scheme, a cross-dimension bidirectional joint path planning algorithm is adopted to generate a safe travel path of the unmanned aerial vehicle and the unmanned vehicle under the path distance constraint.

[0020] According to some embodiments, the present application adopts the technical solutions as follows:

[0021] A computer program product comprising a computer program which, when executed by a processor, implements the unmanned aerial vehicle recovery path adaptive planning method based on air-ground cooperation.

[0022] According to some embodiments, the present application adopts the technical solutions as follows:

[0023] A non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the unmanned aerial vehicle recovery path adaptive planning method based on air-ground cooperation.

[0024] According to some embodiments, the present application adopts the technical solutions as follows:

[0025] An electronic device comprising a processor, a memory, and a computer program, wherein the processor is connected to the memory, and the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the unmanned aerial vehicle recovery path adaptive planning method based on air-ground cooperation.

[0026] Compared with the prior art, the present application has the following beneficial effects:

[0027] The present application constructs a dynamic decision mechanism for unmanned aerial vehicle recovery tasks according to the positions of unmanned aerial vehicles and unmanned vehicles, and the power (maximum flight distance) of the unmanned aerial vehicles, and realizes adaptive allocation of responsibilities of the unmanned aerial vehicles and unmanned vehicles in the recovery tasks. The mechanism fully excavates and utilizes the overall efficiency of the air-ground cooperation system, significantly improves the intelligent level of the cooperative operation and the reliability of the task execution, and effectively guarantees the safe and efficient recovery of the unmanned aerial vehicles under the power constraint condition.

[0028] The present application proposes three unmanned aerial vehicle recovery schemes of Ego, SemiAid, and FullAid, which can realize the safe recovery of the unmanned aerial vehicles while strictly meeting the remaining power constraint, and minimize the influence of the recovery process on the original task of the unmanned vehicle. Specifically, the consideration of the maximum flight distance of the unmanned aerial vehicle is introduced into the unmanned aerial vehicle recovery decision mechanism and the path planning algorithm, the ability of the unmanned aerial vehicle to complete the safe recovery task alone is evaluated, the assistance of the unmanned vehicle is increased only when necessary, and the flight distance of the unmanned aerial vehicle is maximized under the limited power constraint to reduce the degree of deviation of the unmanned vehicle from the original task area, so that the unmanned vehicle can quickly return to the original task area after the completion of the recovery task.

[0029] The application proposes a cross-dimension bidirectional joint path planning algorithm, which can simultaneously maintain two search processes of three-dimensional space forward search and two-dimensional space backward search, and efficiently generate a safe travel path of a UAV and an unmanned vehicle that meets the UAV recovery task. By introducing a guiding item of searching the current node on the other side in the heuristic function, the two search processes can be widely searched while quickly expanding to each other. When the node expands to the safety boundary, the search condition is immediately tightened, and a candidate rendezvous point set is constructed. When selecting the best rendezvous point, the waiting cost evaluation is increased to balance the forward and backward path execution time under the premise of ensuring that the UAV flight distance meets the remaining power constraint, thereby avoiding long waiting time of one side. BRIEF DESCRIPTION OF DRAWINGS

[0030] The accompanying drawings, which form a part of this specification, are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification. The embodiments of these drawings are set to explain the application and do not constitute an improper limitation of the application.

[0031] Figure 1 The flow chart of the UAV recovery path adaptive planning method based on air-ground cooperation of embodiment 1.

[0032] Figure 2 The UAV recovery simulation result graph under the Ego recovery scheme of embodiment 1, wherein (a) is the front view of the simulation result, (b) is the top view of the simulation result, and the blue line is the travel path of the UAV.

[0033] Figure 3 The UAV recovery simulation result graph under the SemiAid recovery scheme of embodiment 1, wherein (a) is the front view of the simulation result, (b) is the top view of the simulation result, the blue line is the travel path of the UAV, and the red line is the travel path of the unmanned vehicle.

[0034] Figure 4 The UAV recovery simulation result graph under the FullAid recovery scheme of embodiment 1, wherein (a) is the front view of the simulation result, (b) is the top view of the simulation result, the blue line is the travel path of the UAV, and the red line is the travel path of the unmanned vehicle. DETAILED DESCRIPTION

[0035] The application will be further described below in conjunction with the drawings and embodiments.

[0036] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as generally understood by those skilled in the art to which the application belongs.

[0037] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0038] Embodiment 1

[0039] In an embodiment of the present application, a method for adaptive planning of a UAV recovery path based on air-ground cooperation is provided, comprising:

[0040] Step one: construct a three-dimensional grid map of the task environment, obtain the positions of the UAV and the unmanned vehicle, and the maximum flight distance of the UAV;

[0041] Step two: based on the positions of the UAV and the unmanned vehicle and the maximum flight distance of the UAV, determine the responsibility allocation of the UAV and the unmanned vehicle in the recovery task, and obtain the best UAV recovery scheme;

[0042] Step three: according to the best UAV recovery scheme, select the corresponding path planning mode, and determine the starting point, ending point and dimension of the search space of the path planning based on the positions of the UAV and the unmanned vehicle;

[0043] Step four: according to the selected path planning mode, execute the path planning algorithm, and generate the final recovery path;

[0044] Wherein, when the UAV and the unmanned vehicle respectively undertake part of the recovery responsibility in the recovery scheme, a cross-dimension bidirectional joint path planning algorithm is used to generate a safe travel path of the UAV and the unmanned vehicle under the constraint of path distance.

[0045] The responsibility distribution herein is to clearly distribute the responsibilities to be borne by the UAV and the unmanned vehicle in the recovery process, which is based on the actual state of the system (the positions of the UAV and the unmanned vehicle, and the maximum flight distance of the UAV). The responsibility distribution is the key to the air-ground cooperation to achieve the recovery task and is the basis for forming the UAV recovery scheme. The responsibilities borne by the UAV and the unmanned vehicle are different in the three recovery schemes, and the path planning algorithms adopted are also different: in the Ego recovery scheme, the unmanned vehicle only provides a landing platform, the recovery responsibility is completely distributed to the UAV, only the path of the UAV needs to be planned (the unmanned vehicle executes the original task, and the path planning thereof is not within the scope of the present application), and a one-way path planning algorithm is adopted; in the SemiAid recovery scheme, the recovery responsibility is distributed to the UAV and the unmanned vehicle, and the paths of the UAV and the unmanned vehicle need to be planned simultaneously, a cross-dimension bidirectional joint path planning algorithm is adopted; and in the FullAid recovery scheme, the recovery responsibility is mainly distributed to the unmanned vehicle, and the UAV only cooperates with the unmanned vehicle to complete the landing after the unmanned vehicle arrives, only the path of the unmanned vehicle needs to be planned, and a one-way path planning algorithm is adopted.

[0046] As an embodiment, the UAV recovery path adaptive planning method based on air-ground cooperation of the present application realizes safe and efficient recovery of the UAV under the power constraint through the air-ground cooperative decision and the cross-dimension bidirectional joint path planning algorithm, significantly improves the task success rate and the system cooperation efficiency, and the specific implementation process is described below.

[0047] S1 initialization: a three-dimensional grid map of the task environment is constructed, and the positions of the UAV and the unmanned vehicle and the maximum flight distance of the UAV are obtained;

[0048] Further, the positions of the UAV and the unmanned vehicle are provided by a positioning system, and the maximum flight distance of the UAV can reflect the remaining power of the battery, which is calculated according to the remaining power of the UAV.

[0049] Specifically, the grid map in the present embodiment is a three-dimensional grid map, and the grid resolution in each dimension is 1m. Figure Three The coordinate origin is (0, 0, 0). The actual position information of the UAV is obtained by an optical motion capture positioning system. The actual position information of the unmanned vehicle is obtained by an optical motion capture positioning system. The maximum flight distance of the UAV is 1000m.

[0050] S2 UAV recovery scheme decision: based on the positions of the UAV and the unmanned vehicle and the maximum flight distance of the UAV, the responsibility distribution of the UAV and the unmanned vehicle in the recovery task is determined, and the best UAV recovery scheme is obtained;

[0051] The optional UAV recovery schemes include Ego, SemiAid and FullAid, and specifically are:

[0052] ​​(1) Ego recovery scheme in which the UAV alone assumes the responsibility of recovery: the UAV navigates to the unmanned vehicle landing platform and lands by itself, and the task of the unmanned vehicle is not affected during the entire recovery process;

[0053] (2) SemiAid recovery scheme in which the UAV and the unmanned vehicle assume part of the recovery responsibility: the unmanned vehicle suspends the task it is performing, approaches the UAV, and under the guidance of the optimal meeting point, the UAV lands on the landing platform of the unmanned vehicle to complete the recovery task, and then the unmanned vehicle returns to the original task area to continue performing the task;

[0054] (3) FullAid recovery scheme in which the unmanned vehicle assumes the main recovery responsibility: the UAV starts the power saving mode or temporarily lands and waits for the unmanned vehicle, and the unmanned vehicle suspends the task it is performing and goes to the location of the UAV to assist the UAV in completing the landing.

[0055] Specifically, the optimal UAV recovery scheme is selected according to the following formula:

[0056]

[0057] wherein are normal numbers, respectively representing a self-defined safety threshold and a distance coefficient, represents the distance between the UAV and the unmanned vehicle.

[0058] In the FullAid recovery scheme, the UAV is seriously short of power, and in the power saving mode, the UAV will turn off the power of unnecessary hardware (such as lights, thermal imagers, etc.), and the UAV will hover at the minimum power required for flight, waiting for the unmanned vehicle; since the power of the UAV in the power saving mode is still being consumed, in order to further ensure the safety of the UAV and improve the possibility of successful recovery, the UAV can also be directly landed in a nearby optional landing area (to avoid crashing due to power depletion), and after the unmanned vehicle arrives, the UAV can take off and land on the landing platform of the unmanned vehicle, or be moved to the landing platform of the unmanned vehicle by the clamping device on the unmanned vehicle; as for whether to choose to hover in the power saving mode or to temporarily land, it can be selected according to the actual situation of the UAV.

[0059] The main recovery responsibility here is not equivalent to assuming the responsibility of recovery alone. In the FullAid recovery scheme, the distance between the UAV and the unmanned vehicle is shortened by the unmanned vehicle, and the main recovery responsibility is assumed by the unmanned vehicle; after the unmanned vehicle arrives near the location of the UAV, the UAV needs to cooperate with the unmanned vehicle to complete the landing, that is, in the landing stage, the UAV needs to assume a certain recovery responsibility; in the Ego recovery scheme, the cooperation of the unmanned vehicle is not required during the entire recovery process, and the unmanned vehicle only provides the landing platform, and its original task will not be affected, so the UAV alone assumes the responsibility of recovery.

[0060] S3 path planning mode decision: according to the best unmanned aerial vehicle recovery scheme, the corresponding path planning mode is selected, and the starting point, ending point and dimension of the search space of the path planning are determined based on the positions of the unmanned aerial vehicle and the unmanned vehicle;

[0061] Further, the optional path planning mode includes two kinds of one-way and two-way, specifically:

[0062] (1) In the one-way path planning mode, the path planning algorithm has the ability of two-dimensional and three-dimensional path planning, has a determined starting point and ending point, only maintains one search process, and finally generates a path;

[0063] (2) In the two-way path planning mode, the path planning algorithm has two determined starting points, and simultaneously maintains two search processes of different dimensions and different directions: forward search and backward search. The two search processes meet at the best meeting point determined in the search process, and finally two paths are generated: the three-dimensional flight path of the unmanned aerial vehicle and the two-dimensional travel path of the unmanned vehicle.

[0064] Further, different recovery schemes correspond to different path planning modes, and the starting point, ending point and dimension of the search space of the planning algorithm are also different, as follows:

[0065] The ego recovery scheme corresponds to the one-way path planning mode, the starting point of the planning algorithm is the position of the unmanned aerial vehicle , the ending point is the position of the unmanned vehicle , and the dimension of the search space is ;

[0066] The SemiAid recovery scheme corresponds to the two-way path planning mode, the forward search starting point of the planning algorithm is the position of the unmanned aerial vehicle , the dimension of the search space is , the backward search starting point is the position of the unmanned vehicle , the dimension of the search space is , and the ending points of the forward search and the backward search are the same, which are the best meeting points optimized by comprehensively considering the flight distance of the unmanned aerial vehicle, search efficiency and the influence degree of the original task of the unmanned vehicle and other factors in the search process;

[0067] The FullAid recovery scheme corresponds to the one-way path planning mode, the starting point of the planning algorithm is the position of the unmanned vehicle , the ending point is the projection position of the unmanned aerial vehicle on the plane where the unmanned vehicle is located , and the dimension of the search space is .

[0068] S4 path planning: according to the selected path planning mode, execute the path planning algorithm to generate the final recovery path. If the generated path meets the unmanned aerial vehicle recovery task requirement, the process ends; otherwise, execute S2.

[0069] If the recovery scheme determined in step S2 is Ego or FullAid, a one-way path planning mode is adopted in step S3, and then the one-way path planning algorithm is used, and the specific steps are as follows:

[0070] SS401, initialize the open list, the closed list and the initial node;

[0071] Specifically, the rasterized coordinates of the start point and the end point are , wherein is the floor operator; the open list is , the closed list is , and the initial node coordinates are , the actual cost is , and the heuristic cost is:

[0072]

[0073] wherein respectively represent the minimum, medium and maximum values in the absolute values of each element.

[0074] The comprehensive cost of the initial node is:

[0075]

[0076] wherein is a weight coefficient.

[0077] The initial node is inserted into the open list .

[0078] SS402, construct a corresponding node increment vector based on the spatial dimension;

[0079] Specifically, the node increment vector represents the coordinate increment and the cost consumed when the node is expanded by one step in each direction; when searching in a two-dimensional space, the node can only be expanded in eight directions (axial and 2D diagonal) on the plane, so the node increment vector in a two-dimensional space is wherein represents the node increment, which is composed of the coordinate increment and the actual cost increment .

[0080] When searching in a three-dimensional space, the node can be expanded in 26 directions (axial, 2D diagonal and 3D diagonal), so the node increment vector in a three-dimensional space is wherein represents the node increment, which is composed of the coordinate increment ​​and actual cost increment Two parts, .

[0081] SS403, based on the start and end positions, an initial ellipse expansion constraint is constructed to limit the expansion range of the node in the horizontal direction;

[0082] Specifically, the ellipse expansion constraint is obtained by rotating and scaling a unit circle, which is expressed in formula as:

[0083]

[0084] wherein, is a point in a circle with a center at the origin and a radius of 1, is a rotation matrix, is a scaling matrix, is an ellipse with a length of the major axis, is an ellipse with a length of the minor axis, is the center of the ellipse .

[0085] The center of the initial ellipse expansion constraint is set as the midpoint of the two-dimensional projection line segment connecting the start and end points, and the scaling matrix is set according to the length of the line segment. Specifically, the center , the length of the major axis , the length of the minor axis , , of the initial ellipse expansion constraint are respectively the coordinate values of the start point , and the coordinate values of the end point .

[0086] During the search process, the size of the ellipse is dynamically adjusted according to the search state. The introduction of the ellipse expansion constraint limits the search to the vicinity of the line segment connecting the start and end points, avoiding a large amount of meaningless search.

[0087] SS404, if the open list is not empty, step SS405 is performed, otherwise, step SS406 is performed;

[0088] SS405, the node with the minimum comprehensive cost in the open list is extracted as the current node, if its coordinates are consistent with the end point coordinates, the safe travel path is obtained by node backtracking, and the algorithm ends; otherwise, the current node is taken as the parent node, a new node to be expanded is constructed based on the node increment vector, the actual cost, the heuristic cost and the comprehensive cost of the effective node to be expanded are calculated, and it is inserted into the open list , execute step SS404;

[0089] Specifically, the current node coordinate is , i.e. the node with the minimum comprehensive cost in is taken out and inserted into the closed list , and deleted in .

[0090] If the current node coordinate is inconsistent with the end point coordinate , the current node is superimposed with the node increment to construct a new node to be expanded, whose coordinate is , and the actual cost is . From the constructed node to be expanded, the node that meets the elliptical expansion constraint, the safety constraint (the node is in the safety area, i.e. the obstacle-free area) and is not in is selected as an effective node to be expanded, and after updating its parent node and cost or calculating its cost, it is inserted into the open list .

[0091] Specifically, if the effective node to be expanded is already in , its actual cost and comprehensive cost are updated, and its parent node is set to the current node, otherwise, its heuristic cost is calculated:

[0092]

[0093] wherein respectively represent the minimum, medium and maximum values in the absolute values of each element of , and is the angle formed by the new node and the start point and the end point. This function considers the axial, 2D diagonal and 3D diagonal movement costs, and the estimate given is closer to the real path cost, which helps to guide the search to approach the target point faster; at the same time, the path bending degree penalty term is introduced to encourage the algorithm to reduce unnecessary turning and improve path quality. The comprehensive cost of the effective node to be expanded is calculated as:

[0094]

[0095] SS406, if the current search boundary is the maximum search boundary, the search fails, and the algorithm ends; otherwise, the size of the ellipse is moderately increased to expand the search range, and step SS404 is executed.

[0096] Specifically, the center of the ellipse remains unchanged, and the size of the ellipse is adjusted by the scaling matrix, i.e. the major axis and the minor axis are simultaneously expanded to times, ,in This is the scaling adjustment factor.

[0097] If the recycling scheme determined in step S2 is SemiAid, then the bidirectional path planning mode is adopted in S3. The corresponding cross-dimensional bidirectional joint path planning algorithm has the following specific steps:

[0098] SD401 initializes the open list, closed list, and initial node for the forward and backward searches;

[0099] Specifically, the rasterized coordinates of the starting points for the forward and backward searches are respectively... , ,in This is the floor operation. The open lists for the forward search and backward search are respectively... , The closed lists are respectively , The initial node coordinates are respectively , The actual costs are respectively , The heuristic costs are as follows:

[0100]

[0101]

[0102] in, They represent The minimum, medium, and maximum absolute values ​​of each element. They represent The combined costs for the minimum, medium, and maximum absolute values ​​of each element are as follows:

[0103]

[0104]

[0105] in, These are the weighting coefficients.

[0106] Will insert ,Will insert .

[0107] SD402 simultaneously constructs both three-dimensional spatial node increment vectors and two-dimensional spatial node increment vectors;

[0108] Specifically, the node increment vector in the forward search of the three-dimensional space. Increment of each node including coordinate increments and actual cost increments two parts, .

[0109] node increment vectors of a two-dimensional space searched backward each node increment including coordinate increments and actual cost increments two parts, .

[0110] SD403, based on the forward search starting point and the backward search starting point, constructs an initial elliptical expansion constraint to limit the expansion range of the node in the horizontal direction;

[0111] Specifically, the elliptical expansion constraint is obtained by rotating and scaling a unit circle, which is expressed by the formula:

[0112] ,

[0113] wherein, is a point in a circle with a radius of 1 and a center at the origin, is a rotation matrix, is a scaling matrix, is the length of the major axis of the ellipse , is the length of the minor axis of the ellipse , is the center of the ellipse .

[0114] The center of the initial elliptical expansion constraint is set as the midpoint of the two-dimensional projection line segment connecting the starting point and the ending point, and the scaling matrix is set according to the length of the line segment. Specifically, the center , the length of the major axis , and the length of the minor axis , , of the initial elliptical expansion constraint are respectively the coordinate values of the forward search starting point , and the coordinate values of the backward search starting point .

[0115] During the search process, the size of the ellipse is dynamically adjusted according to the search state. The introduction of the elliptical expansion constraint limits the search near the line segment connecting the starting point and the ending point, avoiding a large amount of meaningless search. In particular, in the bidirectional path planning, both the forward search process and the backward search process are limited within the elliptical region, greatly increasing the probability of the two search processes meeting.

[0116] SD404: If the forward search has ended, proceed to step SD407; otherwise, extract the node with the minimum overall cost from the forward open list as the current node, and construct a set of forward candidate rendezvous points based on the maximum flight distance of the UAV.

[0117] Specifically, the coordinates of the current node in the forward search are: That is, forward open list The node with the lowest overall cost has an actual cost of The node from Remove and insert into the forward closed list and in Delete. If Then add the current node to the forward candidate rendezvous point set. .

[0118] SD405: If the coordinates of the current node are consistent with the coordinates of the starting point of the backward search, the forward search process ends; otherwise, a new node to be expanded is constructed based on the current node and the incremental vector of the nodes in the three-dimensional space. The actual cost, heuristic cost and comprehensive cost of the effective node to be expanded are calculated and inserted into the open list.

[0119] Specifically, the coordinates of the current node Coordinates of the starting point of the backward search If there is a discrepancy, a new node to be expanded is constructed by overlaying the current node with the incremental nodes in the three-dimensional space, and its coordinates are... The actual cost is From the constructed nodes to be expanded, select those that simultaneously satisfy the elliptical expansion constraint, the safety constraint (the node is within the safe region, i.e., the obstacle-free region), and the flight distance constraint, and are not located in the area specified in the original text. The node in the list is treated as a valid node to be expanded, and after updating its parent node and cost, or calculating its cost, it is inserted into the forward open list. .

[0120] Specifically, if the valid node to be expanded is already... If the current node is selected, its actual cost and total cost are updated, and its parent node is set to the current node; otherwise, its heuristic cost is calculated.

[0121] ,

[0122]

[0123] in, They represent The minimum, medium, and maximum absolute values ​​of each element. Let this be the angle formed by the new node and the starting points of both the forward and backward searches. Calculate the comprehensive cost of the effective nodes to be expanded:

[0124]

[0125] Determine the forward search boundary based on the maximum flight distance of the UAV. and relative search boundaries , The flight distance constraint is determined by the actual cost of the new node. If the node is not found, it is discarded. To improve the speed of finding the optimal meeting point, after the size of the forward candidate meeting point set reaches a custom threshold, new nodes close to the starting point are filtered out, restricting new nodes to the neighborhood before the forward search boundary. .

[0126] SD406: If the forward open list is empty, increase the size of the ellipse appropriately within the maximum search range. If the current search range has reached the boundary of the maximum search range, end the forward search process.

[0127] The ellipse size adjustment is the same as in step SS406, and will not be repeated here.

[0128] SD407: If the backward search has ended but the forward search has not ended, proceed to step SD410; if both the backward search and the forward search have ended, proceed to step SD411; otherwise, extract the node with the minimum comprehensive cost from the backward open list as the current node and construct a set of backward candidate meeting points.

[0129] Specifically, the coordinates of the current node in the backward search are: Backward open list The node with the lowest overall cost has an actual cost of ,like Then add the current node to the set of backward candidate rendezvous points. .

[0130] SD408: If the coordinates of the current node are consistent with the two-dimensional projection coordinates of the starting point of the forward search, then the forward search process ends; otherwise, a new node to be expanded is constructed based on the current node and the incremental vector of the node in the two-dimensional space. The actual cost, heuristic cost and comprehensive cost of the effective node to be expanded are calculated and inserted into the open list.

[0131] Specifically, if the current node coordinates Two-dimensional projection coordinates of the forward search starting point If they are inconsistent, a new node to be expanded is constructed by superimposing the current node with the incremental nodes in the two-dimensional space, and its coordinates are... The actual cost is From the constructed nodes to be expanded, select those that simultaneously satisfy the elliptic expansion constraint and the safety constraint (nodes are within the safe region, i.e., the obstacle-free region) and are not located in the area specified in the original text. The node in the list is treated as a valid node to be expanded, and after updating its parent node and cost, or calculating its cost, it is inserted into the back open list. .

[0132] Specifically, if the valid node to be expanded is already... If the current node is selected, its actual cost and total cost are updated, and its parent node is set to the current node; otherwise, its heuristic cost is calculated.

[0133]

[0134]

[0135] in, They represent The minimum, medium, and maximum absolute values ​​of each element. Let this be the angle formed by the new node and the starting points of both the forward and backward searches. Calculate the comprehensive cost of the effective nodes to be expanded:

[0136]

[0137] The relative search boundary for the backward search is determined based on the maximum flight distance of the UAV. , To fully utilize the drone's remaining battery power, minimize the impact on the vehicle's original mission, and simultaneously improve the speed of finding the optimal rendezvous point, after the size of the backward candidate rendezvous point set reaches a custom threshold, new nodes close to the forward and backward search starting points are filtered out, restricting new nodes to the neighborhood of the drone's maximum flight distance. .

[0138] SD409: If the backward open list is empty, the size of the ellipse is increased appropriately within the maximum search range. If the current ellipse has reached the boundary of the maximum search range, the backward search process ends.

[0139] The ellipse size adjustment is the same as in step SS406, and will not be repeated here.

[0140] SD410: Based on the states of the forward and backward searches and the size of the candidate meeting point set, attempt to calculate the intersection of the forward candidate meeting point set and the backward candidate meeting point set to find the optimal meeting point. If the optimal meeting point is found, proceed to step SD411; otherwise, proceed to step SD404.

[0141] Specifically, the actual cost of forward candidate meeting points exist Internal value selection, backward candidate meeting point heuristic cost Traditional Chinese heuristic cost exist The inner value is taken, where The nodes processed in the pre-planning stage are far away from the safety boundary, allowing a wider search. When the nodes expand to the safety boundary, the boundary control mechanism is triggered, the search condition is tightened, and the pre / post candidate rendezvous point set is constructed to improve the speed of finding an effective rendezvous point. The expansion range of the forward search node is determined by the actual cost , and the furthest expansion is to , ensuring that the forward search does not exceed the preset safety distance to strictly limit the flight path of the UAV within the endurance range. The expansion range of the backward search node is determined by the traditional heuristic cost in the backward search heuristic cost , and the furthest expansion is to to reduce unnecessary expansion.

[0142] Further, when the size of the forward and backward candidate rendezvous point sets both reach the threshold, or when one-sided search terminates and the size of the candidate rendezvous point set on the other side reaches the threshold, the candidate rendezvous point set is calculated, where the comprehensive cost of each node is recalculated using the following formula:

[0143]

[0144] where and represent the actual cost obtained in the forward search process and the actual cost obtained in the backward search process respectively, and represent the expected speed of the UAV and the unmanned vehicle respectively, represents the waiting cost weight, represents the waiting time penalty term, which is used to balance the execution time of the forward and backward paths to avoid long waiting time on one side. If the candidate rendezvous point set is not empty, the node with the smallest comprehensive cost in the candidate rendezvous point set is selected as the best rendezvous point.

[0145] SD411, if the best rendezvous point is not found, switch to the FullAid recovery scheme, enable the one-way path planning algorithm, calculate the safe travel path of the unmanned vehicle, and the algorithm ends; otherwise, the safe travel paths of the UAV and the unmanned vehicle are obtained through the forward search node backtracking and the backward search node backtracking, and the algorithm ends.

[0146] Specifically, if the above steps fail to find the best rendezvous point, switch to the FullAid recovery scheme, execute steps SS401-SS406 to obtain the safe travel path of the unmanned vehicle. Otherwise, based on the best rendezvous point, the safe travel paths of the UAV and the unmanned vehicle are obtained respectively in and The process involves node backtracking to obtain the safe travel paths of drones and unmanned vehicles.

[0147] The heuristic cost functions designed in steps SD405 and SD408 respectively differ from those in traditional heuristic cost functions. and Based on this, a guiding term for the current node in the reverse search was added. and This encourages forward and backward searches to meet in the middle region, increasing the probability and speed of the two searches meeting.

[0148] Figures 2-4 Simulation results for three drone recovery schemes are presented. Figure 2 The image shows the simulation results of drone retrieval under the Ego retrieval scheme, where the blue line represents the drone's travel path. As can be seen, the drone moves along the generated path and can land directly on the unmanned vehicle to complete the retrieval task. During this process, the original task of the unmanned vehicle is not affected. Figure 3 The simulation results of drone recovery under the SemiAid recovery scheme are shown. The blue line represents the drone's path and the red line represents the unmanned vehicle's path. It can be seen that the drone and the unmanned vehicle move towards each other along their respective paths. The assistance of the unmanned vehicle effectively shortens the drone's path length, making it possible to safely recover drones with limited remaining power. Figure 4 The simulation results for drone retrieval under the FullAid retrieval scheme are shown. The blue line represents the drone's path, and the red line represents the vehicle's path. It can be seen that the vehicle leaves its original mission point to fully assist the drone in retrieval. Before the vehicle arrives, the drone can switch to power-saving mode or temporarily land in place. When the vehicle approaches, the drone lands on it to complete the retrieval task.

[0149] The main contents of the method proposed in this embodiment are as follows:

[0150] (1) Construct a decision-making mechanism for drone recycling schemes. Based on the remaining power of the drone, determine the responsibility allocation between the drone and the unmanned vehicle in the recycling task. Through effective collaboration between the drone and the unmanned vehicle, ensure that the drone completes the recycling task before its power is exhausted.

[0151] (2) A cross-dimensional bidirectional joint path planning algorithm is proposed. Based on the graph search algorithm, it maintains two search processes in different dimensions and directions at the same time, so as to realize the rapid generation of safe travel paths for UAVs and unmanned vehicles under the power constraints of UAVs.

[0152] (3) An ellipse extension constraint is proposed for unidirectional and bidirectional path planning algorithms to improve the path planning speed. At the same time, the combined search process on the opposite side is guided to accelerate the convergence of the two search processes in bidirectional path planning.

[0153] (4) A bidirectional path execution time balancing mechanism is proposed, which introduces a waiting time penalty term into the comprehensive cost of the candidate rendezvous point to balance the path execution time of the UAV and the unmanned vehicle, avoid long waiting time of one party while meeting the remaining power constraint of the UAV.

[0154] Embodiment 2

[0155] In an embodiment of the present application, an unmanned aerial vehicle recovery path adaptive planning system based on air-ground cooperation is provided, comprising:

[0156] The position acquisition module is configured to construct a three-dimensional grid map of the task environment, and acquire the positions of the unmanned aerial vehicle and the unmanned vehicle and the maximum flight distance of the unmanned aerial vehicle.

[0157] The responsibility allocation module is configured to determine the responsibility allocation of the unmanned aerial vehicle and the unmanned vehicle in the recovery task based on the positions of the unmanned aerial vehicle and the unmanned vehicle and the maximum flight distance of the unmanned aerial vehicle, and obtain the best unmanned aerial vehicle recovery scheme.

[0158] The mode selection module is configured to select a corresponding path planning mode according to the best unmanned aerial vehicle recovery scheme, and determine the starting point, the ending point and the dimension of the search space of the path planning based on the positions of the unmanned aerial vehicle and the unmanned vehicle.

[0159] The path planning module is configured to execute a path planning algorithm according to the selected path planning mode, and generate a final recovery path.

[0160] In the responsibility allocation, a cross-dimension bidirectional joint path planning algorithm is used to generate a safe travel path of the unmanned aerial vehicle and the unmanned vehicle under path distance constraints when the unmanned aerial vehicle and the unmanned vehicle respectively undertake part of the recovery responsibility in the recovery scheme.

[0161] Embodiment 3

[0162] In an embodiment of the present application, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the unmanned aerial vehicle recovery path adaptive planning method based on air-ground cooperation.

[0163] Embodiment 4

[0164] In an embodiment of the present application, a non-transitory computer readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the unmanned aerial vehicle recovery path adaptive planning method based on air-ground cooperation.

[0165] Embodiment 5

[0166] An embodiment of the present application provides an electronic device, comprising a processor, a memory and a computer program; wherein the processor is connected with the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the method for adaptive planning of a UAV recovery path based on air-ground cooperation.

[0167] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate one or more means for implementing the functions specified in the flowcharts and / or block diagrams. Figure One one or more flows and / or blocks Figure One means for carrying out the functions specified in one or more flows and / or blocks.

[0168] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be executed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed by the computer or other programmable data processing apparatus provide one or more means for implementing the functions specified in the flowcharts and / or block diagrams. Figure One one or more flows and / or blocks Figure One steps for carrying out the functions specified in one or more flows and / or blocks.

[0169] The above describes the specific embodiments of the present application in conjunction with the drawings, but is not a limitation on the protection scope of the present application. It should be understood by those skilled in the art that various modifications or changes made on the basis of the technical solutions of the present application without creative labor are still within the protection scope of the present application.

Claims

1. An adaptive planning method for UAV recovery paths based on air-ground cooperation, characterized in that, include: Construct a 3D grid map of the mission environment to obtain the positions of drones and unmanned vehicles, as well as the maximum flight distance of drones; Based on the locations of the drone and the unmanned vehicle (UAV) and the drone's maximum flight distance, the responsibility allocation between the drone and the UAV in the recovery mission is determined, resulting in the optimal drone recovery scheme. Specifically, the best scheme is selected from three options: Ego recovery scheme: The drone assumes the recovery responsibility alone, autonomously navigating to the UAV landing platform and landing, with the UAV's mission unaffected throughout the recovery process; SemiAid recovery scheme: The drone and the UAV each assume partial recovery responsibility. The UAV suspends its current mission and moves towards the drone. Guided by the optimal rendezvous point, the drone lands on the UAV landing platform to complete the recovery mission, and then the UAV returns to its original mission area to continue its mission; FullAid recovery scheme: The UAV assumes the primary recovery responsibility. The drone enters power-saving mode or temporarily lands, waiting for the UAV to pick it up. The UAV suspends its current mission and proceeds to the drone's location to assist in landing. The selection of the optimal solution from the three drone recovery schemes can be expressed by the following formula: in, Both are positive numbers, representing the user-defined security threshold and distance coefficient, respectively. Indicates the distance between the drone and the unmanned vehicle. This represents the maximum flight distance of the drone. Based on the optimal drone recovery scheme, a corresponding path planning mode is selected. The starting point, ending point, and search space dimensions are determined based on the locations of the drone and unmanned vehicle. Specifically, a choice is made between unidirectional and bidirectional path planning modes. The starting point, ending point, and search space dimensions are determined according to the selected mode. The specific scheme is as follows: The Ego recovery scheme adopts a one-way path planning mode, with the starting point being the drone's location and the ending point being the unmanned vehicle's location, and the search space being three-dimensional; SemiAid's recovery solution uses a two-way path planning mode. The forward search starts at the location of the drone and searches in three dimensions, while the backward search starts at the location of the unmanned vehicle and searches in two dimensions. The endpoint is the optimal meeting point determined by multiple factors during the search process. The FullAid recovery scheme selects a one-way path planning mode, with the starting point being the location of the unmanned vehicle and the ending point being the projection position of the drone on the plane where the unmanned vehicle is located. The search space is two-dimensional. The specific steps of the SemiAid recycling scheme are as follows: Step 1: Initialize the open list, closed list, and initial node for the forward and backward search; Step 2: Construct the 3D spatial node increment vector and the 2D spatial node increment vector; Step 3: Based on the forward search start point and the backward search start point, construct the initial ellipse expansion constraint to limit the expansion range of the nodes in the horizontal direction; Step 4: If the forward search has ended, proceed to step 7; otherwise, extract the node with the minimum overall cost from the forward open list as the current node, and construct a set of forward candidate rendezvous points based on the maximum flight distance of the UAV. Step 5: If the coordinates of the current node are consistent with the coordinates of the starting point of the backward search, the forward search process ends; otherwise, a new node to be expanded is constructed based on the current node and the incremental vector of the nodes in the three-dimensional space. The actual cost, heuristic cost and comprehensive cost of the effective node to be expanded are calculated and inserted into the open list. Step 6: If the forward open list is empty, increase the size of the ellipse appropriately within the maximum search range. If the current search range has reached the boundary of the maximum search range, end the forward search process. Step 7: If the backward search has ended but the forward search has not ended, proceed to step 10; if both the backward search and the forward search have ended, proceed to step 11; otherwise, extract the node with the minimum overall cost from the backward open list as the current node and construct a set of backward candidate meeting points. Step 8: If the coordinates of the current node are consistent with the two-dimensional projection coordinates of the starting point of the forward search, then end the backward search process; otherwise, construct a new node to be expanded based on the current node and the two-dimensional spatial node increment vector, calculate the actual cost, heuristic cost and comprehensive cost of the effective node to be expanded, and insert it into the open list. Step 9: If the backward open list is empty, increase the size of the ellipse appropriately within the maximum search range. If the current ellipse has reached the boundary of the maximum search range, end the backward search process. Step 10: Based on the states of the forward and backward searches and the size of the candidate meeting point set, try to calculate the intersection of the forward candidate meeting point set and the backward candidate meeting point set to find the best meeting point. If the best meeting point is found, proceed to step 11; otherwise, proceed to step 4. Step 11: If the best rendezvous point is not found, switch to the FullAid recovery scheme; otherwise, obtain the safe travel path of the drone and the unmanned vehicle by backtracking the forward search nodes and the backward search nodes, and the algorithm ends. Based on the selected path planning mode, the path planning algorithm is executed to generate the final recycling path; In the aforementioned responsibility allocation scheme, when the drone and the unmanned vehicle each assume part of the recovery responsibility, a cross-dimensional bidirectional joint path planning algorithm is used to generate safe travel paths for the drone and the unmanned vehicle under path distance constraints.

2. The adaptive planning method for UAV recovery path based on air-ground cooperation as described in claim 1, characterized in that, The maximum flight distance of the drone is calculated based on the drone's remaining battery power.

3. The adaptive planning method for UAV recovery path based on air-ground cooperation as described in claim 1, characterized in that, The path planning algorithm also includes a unidirectional path planning algorithm used when the UAV assumes the responsibility of recovery alone and the unmanned vehicle assumes the main responsibility of recovery.

4. An adaptive planning system for UAV recovery paths based on air-ground cooperation, employing the adaptive planning method for UAV recovery paths based on air-ground cooperation as described in any one of claims 1-3, characterized in that, include: The location acquisition module is configured to: construct a 3D grid map of the mission environment, acquire the location of the drone and unmanned vehicle, and the maximum flight distance of the drone; The responsibility allocation module is configured to determine the responsibility allocation between the drone and the unmanned vehicle in the recovery task based on the location of the drone and the unmanned vehicle and the maximum flight distance of the drone, so as to obtain the optimal drone recovery plan. The mode selection module is configured to: select the corresponding path planning mode based on the best drone recovery scheme, and determine the starting point, ending point and search space dimensions of the path planning based on the positions of the drone and the unmanned vehicle; The path planning module is configured to execute a path planning algorithm based on the selected path planning mode to generate the final recycling path. In the aforementioned responsibility allocation scheme, when the drone and the unmanned vehicle each assume part of the recovery responsibility, a cross-dimensional bidirectional joint path planning algorithm is used to generate safe travel paths for the drone and the unmanned vehicle under path distance constraints.

5. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the adaptive planning method for UAV recovery path based on air-ground cooperation as described in any one of claims 1-3.

6. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the adaptive planning method for UAV recovery path based on air-ground cooperation as described in any one of claims 1-3.

7. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the adaptive planning method for UAV recovery path based on air-ground cooperation as described in any one of claims 1-3.

Images