Unmanned aerial vehicle cluster cooperative wind field exploration and autonomous energy obtaining flight method

Abstract

The invention relates to the technical field of flight control, in particular to an unmanned aerial vehicle cluster collaborative wind field exploration and autonomous energy obtaining flight method, which comprises the following steps: constructing a flight mission environment of a multi-source composite wind field; and determining the authenticity of the potential wind source area, selecting a target wind source area with the maximum energy obtaining value in the real wind source area, and selecting a corresponding energy obtaining flight strategy based on the wind field type corresponding to the target wind source area to carry out energy obtaining flight. According to the invention, intelligent and efficient response to dynamic wind field information is realized, and energy obtaining flight strategies adapted to different wind fields are accurately triggered, so that the task energy efficiency and task success rate during long endurance are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of flight control technology, in particular to a method for cooperative wind field exploration and autonomous energy acquisition of a UAV cluster. BACKGROUND

[0002] Under the macro background of the rapid development of low-altitude economy, unmanned aerial vehicles (UAVs) as the main carrier of low-altitude flight tasks are increasingly applied in the fields of regional inspection, emergency communication and environmental monitoring. The common feature of these tasks is that they put forward very high requirements for the long-endurance operation capability of UAVs.

[0003] Currently, most medium and small UAVs rely on on-board batteries as the main energy source, and their endurance is limited by the energy density of the batteries and the on-board carrying weight, which makes it difficult to meet the demand for continuous tasks lasting for several hours or even tens of hours. Therefore, how to obtain renewable energy from the flight environment in real time for the replenishment of flight energy has become a core technical approach to breaking through the endurance bottleneck of UAVs and achieving a leap in task capability.

[0004] Wind energy, as the most widely distributed and most abundant natural resource in the atmospheric environment, provides an ideal energy source for UAVs to realize "flying while charging". The flight height of low-altitude UAVs is just in the area where wind energy resources are relatively abundant. By learning the biological principles of albatrosses, vultures and other birds using wind field energy for long-distance and long-time soaring, it is possible for UAVs to capture energy from the environment wind field by performing specific flight maneuvers and flight strategies. Wind field energy mainly exists in two forms that can be utilized by UAVs: (1) vertical wind field energy (updraft): mainly composed of updraft formed by uneven heating of the ground, or topographic lifting air current generated when air flows through obstacles such as mountains. Birds or gliders can efficiently convert the vertical kinetic energy of the updraft into their own gravitational potential energy (height) by continuously circling (static gliding) in the core area of the updraft, thereby obtaining energy net gain. (2) Horizontal wind field energy (gradient wind field): refers to the gradient wind field formed by the change of horizontal wind speed with height. This phenomenon widely exists in the near-surface, sea surface and high-altitude inversion layer due to factors such as surface friction and atmospheric heat conduction. The flight maneuver of albatrosses and other birds to obtain energy from wind gradient is called "dynamic gliding". The basic principle is that in a wind field where wind speed increases with height, the wind's kinetic energy is converted into the kinetic and potential energy of the bird by climbing against the wind, and then the bird glides downwind to further obtain energy, forming a closed-loop energy flight cycle, thereby maintaining the height and achieving long-distance navigation with little or no consumption of its own energy.

[0005] In the field of static gliding, the prior art mainly solves the problem of how to plan the optimal circling climbing trajectory of the unmanned aerial vehicle under the known rising air flow model of the flight environment. The technical core is to establish the gliding performance model of the unmanned aerial vehicle in the circling state, combine it with the rising air flow model, and solve the optimal circling radius, speed and roll angle that can maximize the net power gain at any height, and finally form a complete three-dimensional spiral climbing trajectory that dynamically changes with the height. In the field of dynamic gliding, the prior art focuses on planning a periodic gliding trajectory with maximum energy gain for the unmanned aerial vehicle under the known gradient wind field model of the flight environment. This scheme usually divides a complete gliding period into four stages: upwind climbing, high-altitude turning, downwind descending, and low-altitude turning, and solves the control instruction sequence that can maximize the net energy gain in a single period through optimal control and other theoretical methods.

[0006] Most of the prior art focuses on the optimal flight energy gain strategy planning and design of a single unmanned aerial vehicle in an information complete, single type wind field environment. This framework cannot cope with real scene flight environments with wide area, dynamic and uncertain flight information. The main reasons are as follows: it is difficult for a single unmanned aerial vehicle to efficiently explore, discover and utilize multiple types (such as gradient wind, rising air flow) of randomly distributed wind field energy sources in a wide area; secondly, the unmanned aerial vehicle path planning method is heavily dependent on accurate environmental prior information input. When facing real application scenarios with incomplete flight information and possibly false information, the unmanned aerial vehicle path planning method of the prior art will cause serious flight energy waste and time delay due to the pursuit of false wind source targets; the existing unmanned aerial vehicle cluster coordination strategy often fails to effectively decouple functions and task division for the two completely different task requirements of "wind source exploration" and "flight energy gain", and it is difficult to balance the high energy consumption of "wide area rapid exploration" and the flight goal of "maximum collective energy gain", resulting in low coordination efficiency.

[0007] Therefore, it is necessary to provide a method for unmanned aerial vehicle cluster cooperative wind field exploration and autonomous energy gain flight to solve the above problems. SUMMARY

[0008] In view of the technical problem of how the unmanned aerial vehicle cluster efficiently discovers, identifies and cooperatively utilizes multiple different types of wind field energy under uncertain flight environment wind field conditions to complete the end-to-end long endurance flight, the present application provides a method for unmanned aerial vehicle cluster cooperative wind field exploration and autonomous energy gain flight. Through ingenious task division, asynchronous information sharing and event-driven dynamic decision-making, the present application directs the unmanned aerial vehicle cluster to execute diversified energy gain strategies including multi-segment dynamic gliding and static circling climbing, thereby realizing the global optimization of wind source exploration efficiency and wind field energy gain to solve the existing problems.

[0009] The unmanned aerial vehicle cluster cooperative wind field exploration and autonomous enabling flight method of the present application adopts the following technical scheme, comprising: Constructing a flight task environment of a multi-source composite wind field; The unmanned aerial vehicle cluster is divided into a reconnaissance unit and a decision execution unit, the reconnaissance unit is used to obtain wind source parameters of potential wind source areas in each "reconnaissance stage" in the flight task environment, update a dynamic wind field information library of the unmanned aerial vehicle cluster according to the wind source parameters of the potential wind source areas, and when the potential wind source areas determined according to the wind source parameters are real wind source areas, actively enter the wind source and select a corresponding enabling flight strategy from a preset enabling flight strategy library, wherein the wind source parameters include wind source authenticity, wind source type, wind source intensity and wind source position, and the wind source type includes a gradient wind field and an updraft field;The decision execution unit is used to obtain the wind source enabling value of each real wind source area according to the wind source parameters of each real wind source area in the dynamic wind field information library, take the real wind source area corresponding to the highest wind source enabling value as a target wind source area, select a corresponding enabling flight strategy from the preset enabling flight strategy library according to the wind field type of the target wind source area, and control the unmanned aerial vehicle cluster to perform enabling flight according to the enabling flight strategy.

[0010] Further schemes of the present application are that the step of constructing a flight task environment of a multi-source composite wind field comprises: The multi-source composite wind field includes a gradient wind field dominated by wind shear and an updraft field driven by thermal convection, wherein two layers of uncertainty parameters are introduced for the multi-source composite wind field, the first layer of uncertainty parameters are that before the start of the flight task, the unmanned aerial vehicle cluster obtains initial prior information about potential wind sources in the flight task environment;The second layer of uncertainty parameters are that the position, type and intensity of the real wind source in the potential wind source have random deviations from the initial prior information.

[0011] Further schemes of the present application are that the reconnaissance unit of the unmanned aerial vehicle cluster includes multiple reconnaissance aircrafts, the multiple reconnaissance aircrafts fly to the potential wind source areas in each "reconnaissance stage" allocated in advance, the reconnaissance unit obtains the wind source parameters when entering the potential wind source areas, and updates the dynamic wind field information library of the unmanned aerial vehicle cluster according to the wind source parameters of the potential wind source areas;When the wind source authenticity in the wind source parameters, actively enter the wind source and select a corresponding enabling flight strategy from a preset enabling flight strategy library.

[0012] The further scheme of the present application is that the decision execution unit of the UAV cluster comprises a lead wingman and a following wingman, the lead wingman is used for task allocation, path planning and energy flight strategy of the UAV cluster, the lead wingman is used for obtaining the wind source energy value of each real wind source area when the potential wind source area in the dynamic wind field information base is a real wind source area, taking the real wind source area corresponding to the highest wind source energy value as a target wind source area, selecting the corresponding energy flight strategy from the preset energy flight strategy base according to the wind field type of the target wind source area, and controlling the UAV cluster to perform energy flight according to the energy flight strategy.

[0013] The further scheme of the present application is that the step of taking the real wind source area corresponding to the highest wind source energy value as a target wind source area according to the wind source energy value of each real wind source area is: The potential energy of the UAV using the updraft is taken as the energy value of the updraft, and the energy consumption difference of the aircraft in the normal flight environment and the energy system of the aircraft in the gradient wind field is taken as the energy value of the gradient wind field under the same horizontal flight distance; If the updraft field and the gradient wind field exist simultaneously in the front flight task environment, the wind source area with the maximum energy value is taken as the target wind source area; If at least two updraft fields exist simultaneously in the front flight task environment, the updraft field with the greater energy value is taken as the target wind source area; If at least two gradient wind fields exist simultaneously in the front flight task environment, the gradient wind field with the greater energy value is taken as the target wind source area.

[0014] The further scheme of the present application is that the energy flight strategy in the preset energy flight strategy base comprises a multi-section dynamic gliding strategy corresponding to the gradient wind field and a static hovering and climbing strategy corresponding to the updraft field.

[0015] The further scheme of the present application is that the static hovering and climbing strategy comprises that the UAV cluster performs approach transition trajectories of hovering cut-in point approximation flight, spiral flight trajectories in the updraft field and departure transition trajectories of gliding flight after flying away from the updraft field in turn; The step of obtaining the approach transition trajectory and the departure transition trajectory when the wind source type is the updraft field is: Based on the optimal hovering radius, the optimal roll angle and the optimal hovering airspeed corresponding to the flight trajectory under different flight altitudes, the best cut-in point of the flight trajectory and the flight state parameters corresponding to the best cut-in point are obtained, the flight state parameters comprise the optimal hovering radius, the optimal roll angle and the optimal hovering airspeed at the best cut-in point; Based on the best cut-in point and its corresponding flight state parameters, and using a decoupled three-dimensional path planning algorithm, a geometrically optimal and all-terminal state constraint satisfying approach transition trajectory is generated; A tangent departure strategy is adopted to plan a shortest trajectory planning adjustment trajectory for smoothly transitioning the aircraft from the circling flight state to the shortest trajectory planning adjustment trajectory towards the final target point as the approach transition trajectory.

[0016] Further schemes of the present application are that the step of adopting the tangent departure strategy to plan a shortest trajectory planning adjustment trajectory for smoothly transitioning the aircraft from the circling flight state to the shortest trajectory planning adjustment trajectory towards the final target point as the approach transition trajectory of the approach phase comprises: A tangent line is drawn from the path planning end point and the last circle circling trajectory of the aircraft in the updraft field, and the tangent point of the tangent line and the circle circling trajectory is taken as a circling change-out point; A concentric circle is drawn with the center of the updraft field as the center, the sum of the radius of the last circle circling trajectory and the preset approach safety adjustment distance as the radius, and the intersection point of the tangent line and the concentric circle is taken as the target point of the approach phase; the flight state of the target point is set as the optimal gliding state calculated according to the maximum lift-drag ratio of the aircraft; Based on the circling change-out point and the target point of the approach phase, a three-dimensional path planning algorithm is used to generate an approach transition trajectory of the approach phase.

[0017] Further schemes of the present application are that the multi-segment dynamic gliding strategy comprises a dynamic gliding flight trajectory composed of an approach trajectory, a multi-segment dynamic gliding flight trajectory in the gradient wind field and an approach trajectory; The step of obtaining the dynamic gliding flight trajectory comprises: The start point pose of the multi-segment dynamic gliding flight trajectory of the aircraft in the gradient wind field is taken as a start point target pose, and the end point pose is taken as an end point target pose; In a horizontal two-dimensional plane, the start point pose and the start point target pose of the aircraft are connected, and the end point pose and the end point target pose are connected; by comparing a plurality of "arc-straight line-arc" combinations, a shortest path is selected as an optimal horizontal path; The horizontal flight distance is taken as a reference, and the horizontal flight distance is the optimal horizontal path; an optimal vertical variation profile corresponding to the start point height and climb angle to the target height and climb angle and the end point height and climb angle to the target height and climb angle is planned out; The optimal horizontal path and the optimal vertical variation profile are fused to reconstruct a three-dimensional space kinematic trajectory satisfying all start point and end point state constraints, and the three-dimensional space kinematic trajectory is the dynamic gliding flight trajectory.

[0018] Further, after the current enabling flight strategy task is completely finished, and the front potential wind source area is a false wind source area, the UAV cluster re-switches back to a cruising waiting state, and re-evaluates the latest wind source information in the dynamic wind field information library at the next moment, to decide the next stage of enabling flight.

[0019] The present application has the following advantages: The method of the present application can enable the UAV cluster to realize autonomous, dynamic and intelligent task planning and execution in a complex multi-source wind field environment with significant uncertainty. Through the task division of "sacrificing local energy efficiency of part of the individual (scout) to exchange for the global information advantage of the cluster, to maximize the energy benefit of the subject (wingman formation)", the UAV cluster not only successfully avoids the interference of invalid wind field information, but more importantly, can continuously guide itself to the optimal wind source under global information, thereby verifying the effectiveness of the method and strategy of the present application in improving the energy efficiency and task success rate of long-endurance tasks; secondly, the decision execution unit designs a limited state machine mechanism of "cruising waiting" and "enabling execution". The state transition is not based on a preset script, but is asynchronously triggered by the wind field parameters updated by the scout unit in the shared dynamic wind field information library, thereby realizing intelligent and efficient response to dynamic wind field information, and accurately triggering the enabling flight strategy adapted to different wind fields. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0021] Figure 1 A flowchart of a UAV cluster cooperative wind field exploration and autonomous enabling flight method of the present application; Figure 2 A flight task environment diagram of a multi-source composite wind field of a UAV cluster in an embodiment of the present application; Figure 3 A diagram of the first stage of the wingman executing enabling flight in an embodiment of the present application; Figure 4 A diagram of the second stage of the wingman executing enabling flight in an embodiment of the present application; Figure 5 A diagram of the task trajectory of the scout and the wingman in an embodiment of the present application; Figure 6 A diagram of the take-off flight trajectory planning of the aircraft in the ascending air current field in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work are within the protection scope of the present application.

[0023] An embodiment of a method for wind field exploration and autonomous energy acquisition flight of a UAV cluster according to the present application is shown in Figure 1 , and includes: S1, constructing a flight task environment of a multi-source composite wind field; For example, in a specific embodiment, the present application first constructs a flight task environment of a composite wind field facing a real application scenario, which contains heterogeneity, multi-source and uncertainty. In the wide-area flight space of the UAV, a plurality of wind sources of different physical causes are randomly distributed, including gradient wind field dominated by wind shear and updraft field driven by thermal convection. More importantly, in order to reflect the incompleteness of wind field information prediction in a real flight environment, the wind field environment established in this embodiment introduces two layers of uncertainty parameters: first, before the start of the flight task, the UAV cluster can only obtain initial prior information about the "potential wind source (PWZ)" in the flight area, and the second layer of uncertainty parameters is that the position, type and intensity of the real wind source in the potential wind source have random deviations from the initial prior information.

[0024] For example, in a specific embodiment, the flight task environment is modeled: the task requires the UAV cluster to fly from a total starting point to a total ending point, and in this flight process, the entire task airspace is naturally divided into several non-intersecting "reconnaissance stages" by the geographical distribution of all potential wind source areas. Based on this, the flight task environment of the multi-source composite wind field shown in Figure 2 is established.

[0025] S2, determining the authenticity of the potential wind source area, and selecting a target wind source area with the greatest energy acquisition value in the real wind source area, and selecting a corresponding energy acquisition flight strategy based on the wind field type corresponding to the target wind source area to perform energy acquisition flight; Specifically, to balance wind source exploration efficiency and flight energy consumption during flight, this embodiment dynamically divides the UAV swarm into a reconnaissance unit and a decision execution unit. The reconnaissance unit is used to acquire wind source parameters of potential wind source areas in each "reconnaissance phase" of the flight mission environment, update the dynamic wind field information database of the UAV swarm based on the wind source parameters of potential wind source areas, and actively enter the wind source area and select the corresponding energy-gaining flight strategy from the preset energy-gaining flight strategy database when the potential wind source area determined by the wind source parameters is a real wind source area. The wind source parameters include: wind source authenticity, wind source type, wind source intensity, and wind source location. The wind source type includes gradient wind field and updraft field. The decision execution unit is used to select the real wind source area with the highest wind source energy gain value as the target wind source area based on the wind source energy gain value of each real wind source area, select the corresponding energy-gaining flight strategy from the preset energy-gaining flight strategy database based on the wind field type of the target wind source area, and control the UAV swarm to perform energy-gaining flight according to the energy-gaining flight strategy.

[0026] For example, in one specific embodiment, the reconnaissance unit of the UAV swarm needs to fly at high speed. The core task of the reconnaissance unit is to act as a "sensor" for wind source detection of the UAV swarm. The reconnaissance unit includes multiple reconnaissance aircraft. Based on the known potential wind source areas in the flight mission environment, the multiple reconnaissance aircraft fly towards the potential wind source areas pre-assigned to each "reconnaissance phase" to quickly verify the authenticity of the potential wind source areas. When the reconnaissance unit enters the potential wind source area, it acquires wind source parameters and updates the dynamic wind field information database of the UAV swarm based on the wind source parameters of the potential wind source area. The authenticity, type (updraft field and gradient wind field), intensity, and location of the wind source to be detected are fed back to the UAV swarm to update the dynamic wind field information database of the UAV swarm. At the same time, the reconnaissance unit itself is also endowed with the ability of "opportunistic energy acquisition". When the potential wind source area is confirmed to be a real wind source, it can immediately and proactively enter the wind source area and execute the corresponding energy acquisition flight strategy to replenish its own energy consumption. When the reconnaissance aircraft does not have a reconnaissance mission to perform, the flight speed is switched to a speed cruise that consumes less energy. In this embodiment, the navigator and wingman cruise at a speed of 12 m / s.

[0027] In one specific embodiment, the reconnaissance unit collects wind source parameters of the potential wind source area through its own sensors after reaching the corresponding potential wind source area, and updates the collected wind source parameters to a dynamic wind field information database accessible by all drones in real time, serving as the only and reliable basis for subsequent collaborative decision-making. The dynamic wind field information database is characterized by pre-set enabling flight strategies applicable to the entire process of the two types of core wind fields in nature (gradient wind field and updraft). When making a cluster decision, the appropriate enabling flight strategy can be accurately called from the dynamic wind field information database according to the wind source parameters returned by the reconnaissance drone: namely, a multi-section dynamic gliding strategy is executed in the gradient wind field, and a static hovering and climbing strategy is executed in the updraft.

[0028] In one specific embodiment, the decision execution unit of the drone cluster includes a lead drone and a follow drone. The lead drone is the decision core and main body of the cluster, and the follow drone adopts a flight mode following the lead drone to jointly form the "execution main body" of the cluster. The lead drone is used for task allocation, path planning, and enabling flight strategy of the drone cluster. When the potential wind source area in the dynamic wind field information database is a real wind source area, the lead drone obtains the wind source enabling value of each real wind source area, takes the real wind source area corresponding to the highest wind source enabling value as the target wind source area, selects the corresponding enabling flight strategy from the pre-set enabling flight strategy library according to the type of the target wind source area, and controls the drone cluster to perform enabling flight according to the enabling flight strategy.

[0029] In this embodiment, the default state of the lead drone is a cruising waiting state. In the cruising waiting state, the lead drone leads the cluster to perform economic cruising towards the task final destination, and continuously listens to the updates of the dynamic wind field information database. Its decision logic is as follows: when the dynamic wind field information database is updated to show that a potential wind source area is confirmed as a real target wind source area, the lead drone immediately performs enabling value evaluation according to the type and parameters of the target wind source area, and follows the "high value first" principle, i.e., when a target wind source area with the highest enabling value appears, the lead drone immediately triggers decision-making, interrupts the cruising waiting state, and performs enabling flight according to the corresponding wind field type of the target wind source area from the pre-set enabling flight information library.

[0030] In one specific embodiment, the UAV utilizes the potential energy of the updraft to increase the energy harvesting value of the updraft; the energy consumption difference between the energy system of the aircraft when flying in the normal flight environment and when flying in the gradient wind field is taken as the energy harvesting value of the gradient wind field in the same horizontal flight distance; if both the updraft field and the gradient wind field exist in the front flight task environment, the wind source area with the largest energy harvesting value is taken as the target wind source area; if at least two updraft fields exist in the front flight task environment, the updraft field with the larger energy harvesting value is taken as the target wind source area; if at least two gradient wind fields exist in the front flight task environment, the gradient wind field with the larger energy harvesting value is taken as the target wind source area. The evaluation model of the energy consumption of the energy system of the aircraft when flying in the gradient wind field is as follows:

[0031]

[0032] In the formula, represents the energy consumption of the energy system of the aircraft when flying in the gradient wind field; represents the output power of the energy system of the aircraft when flying in the gradient wind field; represents the output voltage of the energy system; represents the output current of the energy system; represents the input voltage of the electronic speed controller; represents the input current of the electronic speed controller; represents the input voltage of the motor; represents the input current of the motor; represents the internal resistance of the electronic speed controller; represents the correction coefficient; represents the initial time of integration; represents the end time of integration, wherein, when the flight phase is the unpowered gliding flight phase ; the energy consumption of the energy system of the aircraft when flying in the normal flight environment is automatically obtained by the system.

[0033] In one specific embodiment, the energy harvesting flight strategy in the preset energy harvesting flight strategy library includes a multi-section dynamic gliding strategy corresponding to the gradient wind field and a static hovering and climbing strategy corresponding to the updraft field.

[0034] Specifically, in the embodiment, the static spiral climb strategy includes: the UAV cluster sequentially performs a spiral approach flight trajectory for approaching a cut-in point, a spiral flight trajectory in the updraft field, and a spiral exit flight trajectory after flying away from the updraft field, wherein the spiral flight trajectory is obtained by: constructing an updraft model describing physical characteristics of the thermal updraft, obtaining a vertical updraft speed at a spiral radius at any flight height by using the updraft model; constructing a safety constraint condition based on a roll angle range and a stall speed, and constructing a flight performance model of an aircraft unpowered gliding performance considering a spiral maneuver based on the safety constraint condition; obtaining a potential energy power obtained by the aircraft from the updraft based on a gravity and the vertical updraft speed when the aircraft flies in the updraft; obtaining a resistance dissipation power dissipated by the aircraft itself to overcome resistance based on a spiral sinking rate when the aircraft unpowered spirals; obtaining a net energy gain power of the aircraft based on the potential energy power and the resistance dissipation power; taking the spiral radius corresponding to the maximum net energy gain power of the aircraft at different flight heights as the optimal spiral radius at different flight heights, and obtaining the optimal roll angle and the optimal spiral airspeed corresponding to the optimal spiral radius of the aircraft at different flight heights by using the flight performance model; obtaining the spiral climb trajectory based on the optimal spiral radius, the optimal roll angle and the optimal spiral airspeed corresponding to different flight heights, and taking the spiral climb trajectory as the flight trajectory in the updraft. Wherein, the steps of obtaining the approach transition trajectory and the exit transition trajectory when the wind source type is the updraft field are: obtaining the best cut-in point and the flight state parameters corresponding to the best cut-in point based on the optimal spiral radius, the optimal roll angle and the optimal spiral airspeed corresponding to the flight trajectory at different flight heights, the flight state parameters including: the optimal spiral radius, the optimal roll angle and the optimal spiral airspeed at the best cut-in point; generating an approach transition trajectory which is geometrically optimal and meets all terminal state constraints by using a decoupled three-dimensional path planning path algorithm based on the best cut-in point and the flight state parameters corresponding thereto; planning a shortest path planning adjustment trajectory for smoothly transitioning the aircraft from the spiral flight state to the final target point as the exit transition trajectory by using the tangent departure strategy. In the embodiment, as shown in Figure 6 , the steps of planning a shortest path planning adjustment trajectory for smoothly transitioning the aircraft from the spiral flight state to the final target point as the exit transition trajectory by using the tangent departure strategy in the embodiment are: drawing a tangent line from the path planning end point (i.e. Figure 6 , the path planning target point) and the last circle spiral trajectory of the aircraft in the updraft, and taking the tangent point of the circle spiral trajectory as a spiral exit point (i.e. Figure 6The center of the updraft is taken as the center of a concentric circle with a radius being the sum of the radius of the last circle trajectory and a preset take-off safety adjustment distance, and a line connecting the take-off end point and the center of the updraft is drawn, and the intersection point of the line and the concentric circle is taken as a target point of the take-off stage; the flight state of the target point is set as an optimal gliding state calculated according to the maximum lift-drag ratio of the aircraft; and a take-off transition trajectory is generated based on the circle change-out point, the target point of the take-off stage and by using a three-dimensional path planning algorithm.

[0035] Specifically, in the embodiment, the multi-segment dynamic gliding strategy includes a dynamic gliding flight trajectory composed of an approach trajectory, a multi-segment dynamic gliding flight trajectory in a gradient wind field and a take-off trajectory; the multi-segment dynamic gliding flight trajectory in the gradient wind field adopts the multi-segment dynamic gliding flight trajectory in the gradient wind field disclosed in the patent with the application number 2025116401081, wherein the step of obtaining the dynamic gliding flight trajectory is: taking the flight trajectory start point pose of the multi-segment dynamic gliding flight trajectory of the aircraft in the gradient wind field as a start point target pose and taking the flight trajectory end point pose as an end point target pose; in a horizontal two-dimensional plane, the start point pose and the start point target pose of the aircraft are connected, and the end point pose and the end point target pose are connected, and by comparing a plurality of "arc-straight line-arc" combinations, a path with the shortest total length is selected as an optimal horizontal path; the horizontal flight distance is taken as a reference, the horizontal flight distance is the optimal horizontal path, and the optimal vertical variation profile corresponding to the start point height and climb angle to the target height and climb angle and the end point height and climb angle to the target height and climb angle is planned; the optimal horizontal path and the optimal vertical variation profile are fused to reconstruct a three-dimensional space kinematic trajectory meeting all start point and end point state constraints, which is the dynamic gliding flight trajectory.

[0036] For example, after the current energy acquisition flight strategy task is completely finished, and the front potential wind source area is a false wind source area, the UAV cluster switches back to the cruising waiting state, and the latest wind source information in the dynamic wind field information library at the next moment is re-evaluated to decide the next stage of energy acquisition flight.

[0037] It should be noted that the task division, asynchronous exploration, information sharing and energy acquisition strategy execution of the UAV cluster are closely combined to form the energy acquisition flight strategy, so that the UAV cluster can avoid the risk brought by wind source uncertainty to the greatest extent, and can flexibly adopt the most efficient wind field energy acquisition mode according to the actual discovered wind field type, thereby realizing efficient and robust cooperative energy acquisition and long flight time in an unknown multi-source wind field environment.

[0038] The application will be further described below in combination with specific embodiments and drawings: Step one, construct the flight task environment of multi-source wind field; The flight task environment of UAV cluster is set as a space cuboid region, the length (Y direction) of the space cuboid region is 19000m, the width (X direction) is 2000m, and the height (Z direction) is 500m.

[0039] Six potential wind source zones (PWZ) are set in the flight region, and the six potential wind source zones are numbered as [1, 2, 3, 4, 5, 6]. The potential wind source includes gradient wind field and updraft. The types and initial prior information of the six potential wind source zones are shown in Table 1. In order to reflect the uncertainty of wind source information in real scene, it is set that there is one false wind source in the six potential wind source zones, and the real wind source parameters of the six wind sources also have random deviation from the initial prior information.

[0040] Table 1

[0041] The UAV cluster is set to be composed of 10 UAVs, and the cruise flight speed of the UAV is 12m / s. The flight task region is given by the task starting point and task ending point parameters. Among the 10 UAVs, 2 UAVs are set as reconnaissance aircraft, and the task is to explore the wind source. The remaining 8 UAVs are composed of 1 leader aircraft and 7 follower aircraft, and the task is to obtain wind energy flight according to the wind source parameter information fed back by the reconnaissance aircraft. The flight performance of the UAV is given according to the real small UAV as shown in Table 2.

[0042] Table 2

[0043] Step two, determine the authenticity of the potential wind source zone, and select the target wind source zone with the maximum energy value in the real wind source zone, and select the corresponding energy flight strategy based on the wind field type corresponding to the target wind source zone to perform energy flight.

[0044] 1. The reconnaissance aircraft performs the wind source search task: According to the initial prior information of the potential wind source given in Table 1, two reconnaissance aircrafts are responsible for the wind source parameter exploration and determination of three potential wind sources respectively.

[0045] The reconnaissance aircraft No. 1 is assigned the exploration task of the PWZ No. [1, 3, 6], and at the flight time t = 38.08s, it first enters the No. 1 potential wind source and confirms that the No. 1 gradient wind field actually exists, and the wind source information of the No. 1 potential wind source is transmitted to the lead wingman. The reconnaissance aircraft No. 1 starts to execute the multi-segment dynamic gliding strategy to obtain the energy of the gradient wind field corresponding to the No. 1 potential wind source. After completing the flight out of the No. 1 potential wind field area, it continues to fly forward. At t = 492.67s, it enters the No. 3 potential wind field and identifies that the No. 3 potential wind field is a “false wind field”, and after the information is transmitted to the lead wingman, it directly flies to the last target, and finally, at t = 888.48s, it enters the No. 6 potential wind field area and verifies that the No. 6 potential wind field is a real wind field; in the No. 6 potential wind field area, it flies according to the predetermined flight strategy until t = 1345.58s to reach the end point. The total energy consumption of the No. 1 reconnaissance aircraft in the whole flight process is 56144.00J.

[0046] The reconnaissance aircraft No. 2 is responsible for exploring the PWZ No. [2, 4, 5], and at t = 54.94s, it enters the No. 2 potential wind field and confirms that the No. 2 updraft type wind field actually exists, and after transmitting the wind source information, the reconnaissance aircraft No. 2 immediately executes the static hovering climbing flight strategy, and after obtaining energy in the No. 2 potential wind field, it continues to execute the next wind source exploration, and at t = 1638.49s, it confirms that the No. 4 potential wind source is a gradient type wind field and actually exists, at this time, it adopts the multi-segment dynamic gliding flight strategy to utilize the energy of the No. 4 wind field, and after obtaining energy, it directly flies to the task end point. Because its task span is large, its total task time is as high as 2643.07s. However, although the flight time is longer, the total energy consumption (27256.36J) is greatly reduced compared with the reconnaissance aircraft No. 1, and the average power is as low as 10.31W. This significant energy efficiency improvement is because it successfully utilizes the energy of the updraft type wind field it discovers in the exploration process, obtains a large amount of gravitational potential energy through the static hovering climbing strategy, and after the subsequent exploration task, switches the cruising speed to the economic cruising speed of the lead wingman, thereby greatly saving the power energy consumption required for subsequent long-distance cruising. This result clearly shows that the exploration and energy acquisition of the method of the present application are not separated, and each unmanned aerial vehicle individual can serve as an energy collection unit of the cluster while completing its main task, thereby maximizing the global energy utilization efficiency of the whole cluster.

[0047] 2. Dynamic decision trajectory of the lead wingman The cluster of the lead wingman and the following wingman is designed to delay take-off for 30 seconds, so as to receive the wind source parameters returned by the No. 1 and No. 2 reconnaissance aircrafts, in order to further decide the cluster flight strategy.

[0048] First stage decision: as Figure 3As shown, although the gradient wind field of wind source No. 1 was identified first, the lead wingman followed the "high value priority" principle and continued to cruise and wait. Until t=54.94s, the higher value updraft wind field No. 2 was confirmed to exist. At this time, the lead wingman aborted its cruise and flew towards wind field No. 2 along its planned path to perform a circling climb. Its trajectory also changed significantly at this moment.

[0049] Second-stage decision-making: such as Figure 4 As shown, during the flight using wind farm 2, wind farm 3 was identified as "false". Therefore, after gaining power from wind farm 2, the lead wingman entered a "cruise waiting" state until t=1638.49s, when wind farm 4 was identified as real, triggering a decision again, causing the cluster to immediately plan a path to wind farm 4.

[0050] The third stage of decision-making: After gaining power at wind field 4, the lead wingman did not undergo any "cruise waiting" and flew directly to wind field 6. This was because wind field 6 had been pre-identified by reconnaissance aircraft number 1. Therefore, after completing its mission at wind field 4, the lead wingman immediately assessed the dynamic wind field information database and found that the information for the next stage was complete, thus achieving a seamless transition between mission stages and ultimately forming the following... Figure 5 The flight path shown.

[0051] 3. Comprehensive task evaluation: As shown in Table 3, Table 3 shows the three-dimensional flight trajectory of the UAV swarm and the collaborative exploration and energy acquisition mission event report in this mission scenario.

[0052] Table 3

[0053] From Table 3 and Figure 5 The results showed that the drone swarm's entire mission lasted a total of 2722.38 seconds (approximately 45 minutes). Through sophisticated coordination strategies, the swarm successfully utilized all available real wind fields and avoided false intelligence, ultimately reaching the mission endpoint.

[0054] It is necessary to note that the comprehensive evaluation of the whole task not only focuses on the total energy consumption, but also deeply analyzes the energy cost and contribution of different roles in the cluster. First, the scout aircraft exchanges valuable confirmation information and decision-making time for the whole cluster at the cost of its own high energy consumption. The average power of the two scout aircrafts is as high as 20.91 W, far exceeding the flight wingman. Their high-speed forward exploration is a necessary energy investment, and the purpose is to reduce the huge risk faced by the whole cluster due to environmental uncertainty. In particular, the average power of the first scout aircraft is as high as 41.72 W, and it makes a key contribution to the cluster in the early stage by bearing more intensive exploration tasks to quickly exclude interference and lock the optimal wind source target. Second, the lead wingman is the biggest energy beneficiary of this cooperative strategy as the decision-making core and execution subject. Benefiting from the accurate intelligence provided by the scout aircraft, the lead wingman can maximize the avoidance of invalid flight and blind waiting. Its trajectory is almost entirely composed of efficient energy acquisition maneuvers and economic cruise flight. Therefore, although its flight time is as long as 2722.38 s, its average power is only 9.61 W, and the total energy consumption is only 26161.19 J, showing a high wind field energy utilization efficiency. Finally, to evaluate the system total energy consumption of the whole 10-aircraft cluster, the consumption of the 7 follower wingmen needs to be counted. Since the follower wingmen completely reproduce the energy-saving trajectory of the lead wingman, their energy consumption is the same as that of the lead wingman. Therefore, the system total energy consumption of the whole cluster is calculated to be 292689.86 J (about 81.30 Wh).

[0055] The case simulation results show that the method can enable the UAV cluster to realize autonomous, dynamic, and intelligent task planning and execution in a complex multi-source wind field environment with significant uncertainty. Through the task division of "sacrificing local energy efficiency of part of individuals (scout aircrafts) to exchange global information advantage of the cluster, thereby maximizing the energy benefit of the main body (wingman formation)", the UAV cluster not only successfully avoids the interference of invalid wind field information, but more importantly, can continuously guide itself to the optimal wind source under global information, thereby verifying the effectiveness of the method strategy in improving the energy efficiency and task success rate of long-haul tasks.

[0056] The above only describes the preferred embodiments of the present application and should not be used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for cooperative wind field exploration and autonomous energized flight of UAV swarm, characterized in that, The application relates to a flight task environment of a multi-source composite wind field. The flight task environment of the multi-source composite wind field comprises the following steps: The multi-source composite wind field comprises a gradient wind field dominated by wind shear and an updraft field driven by thermal convection, wherein two layers of uncertainty parameters are introduced into the multi-source composite wind field, the first layer of uncertainty parameters are initial prior information about potential wind sources in the flight task environment obtained by a UAV cluster before the flight task starts, and the second layer of uncertainty parameters are random deviations of the position, type and intensity of real wind sources in the potential wind sources from the initial prior information.

2. The method of claim 1, wherein, The reconnaissance unit of the UAV cluster comprises a plurality of reconnaissance machines, the plurality of reconnaissance machines fly to potential wind source areas in each "reconnaissance stage" allocated in advance, the reconnaissance unit obtains wind source parameters when entering the potential wind source areas, and the dynamic wind field information database of the UAV cluster is updated according to the wind source parameters of the potential wind source areas; when the wind source real in the wind source parameters, the UAV cluster actively enters the wind source and selects a corresponding energy-gaining flight strategy from a preset energy-gaining flight strategy library. The decision execution unit of the UAV cluster comprises a lead aircraft and a following aircraft, the lead aircraft is used for task allocation, path planning and energy-gaining flight strategy of the UAV cluster, the lead aircraft is used for obtaining the energy-gaining value of each real wind source area when the potential wind source area in the dynamic wind field information database is a real wind source area, taking the real wind source area corresponding to the highest energy-gaining value as a target wind source area, selecting a corresponding energy-gaining flight strategy from a preset energy-gaining flight strategy library according to the wind field type of the target wind source area, and controlling the UAV cluster to perform energy-gaining flight according to the energy-gaining flight strategy. 3.The method of claim 1, wherein, The step of taking the real wind source area corresponding to the highest energy-gaining value as the target wind source area according to the energy-gaining value of each real wind source area comprises the following steps: 4.The method of claim 1, wherein, The potential energy of the UAV using the updraft is taken as the energy-gaining value of the updraft; and the energy consumption difference between the aircraft flying in a normal flight environment and the energy system of the aircraft flying in the gradient wind field is taken as the energy-gaining value of the gradient wind field under the same horizontal flight distance.

5. The method of claim 4, wherein, ​ ​ If the ascending airflow field and the gradient wind field exist simultaneously in the flight task environment in front, the wind source area with the maximum energy value is taken as the target wind source area; If at least two ascending airflow fields exist simultaneously in the flight task environment in front, the ascending airflow field with the greater energy value is taken as the target wind source area; If at least two gradient wind fields exist simultaneously in the flight task environment in front, the gradient wind field with the greater energy value is taken as the target wind source area.

6. The method of claim 1, wherein, The energy-gaining flight strategies in the preset energy-gaining flight strategy library include: a multi-section dynamic gliding strategy corresponding to the gradient wind field and a static hovering and climbing strategy corresponding to the ascending airflow field.

7. The method of claim 6, wherein, The static hovering and climbing strategy includes: the UAV cluster sequentially performs a hovering approach flight trajectory, a spiral flight trajectory in the ascending airflow field, and an out-of-approach transition trajectory after flying away from the ascending airflow field; The steps of obtaining the approach transition trajectory and the out-of-approach transition trajectory when the wind source type is the ascending airflow field are: Based on the optimal hovering radius, the optimal roll angle and the optimal hovering airspeed corresponding to the flight trajectory at different flight altitudes, the best entry point of the flight trajectory and the flight state parameters corresponding to the best entry point are obtained, and the flight state parameters include: the optimal hovering radius, the optimal roll angle and the optimal hovering airspeed at the best entry point; Based on the best entry point and the flight state parameters corresponding thereto, and by using a decoupled three-dimensional path planning algorithm, an approach transition trajectory that is geometrically optimal and meets all terminal state constraints is generated. A tangent departure strategy is adopted to plan a shortest trajectory planning adjustment trajectory that smoothly transitions the aircraft from the hovering flight state to the shortest trajectory planning adjustment trajectory towards the final target point as the out-of-approach transition trajectory.

8. The method of claim 7, wherein, The steps of adopting the tangent departure strategy to plan a shortest trajectory planning adjustment trajectory that smoothly transitions the aircraft from the hovering flight state to the shortest trajectory planning adjustment trajectory towards the final target point as the out-of-approach transition trajectory in the out-of-approach phase are: A tangent line is drawn from the path planning end point and the last circle hovering trajectory of the aircraft, and the tangent point of the tangent line and the circle hovering trajectory is taken as a hovering exit point; A concentric circle is drawn with the center of the ascending airflow field as the center, and the sum of the radius of the last circle hovering trajectory and the preset out-of-approach safety adjustment distance is taken as the radius, and the intersection point of the connecting line between the path planning end point and the center of the ascending airflow field and the concentric circle is taken as the target point in the out-of-approach phase; the flight state of the target point is set to the best gliding state calculated according to the maximum lift-drag ratio of the aircraft; Based on the hovering exit point and the target point in the out-of-approach phase, and by using a three-dimensional path planning algorithm, an out-of-approach transition trajectory in the out-of-approach phase is generated. 9.The method of claim 6, wherein, The multi-section dynamic gliding strategy includes a dynamic gliding flight trajectory composed of an approach trajectory, a multi-section dynamic gliding flight trajectory in the gradient wind field and an out-of-approach trajectory; The steps of obtaining the dynamic gliding flight trajectory are: The start point pose of the multi-section dynamic gliding flight trajectory of the aircraft in the gradient wind field is taken as the start point target pose, and the end point pose is taken as the end point target pose. In the horizontal two-dimensional plane, the start point pose of the aircraft and the start point target pose, the end point pose and the end point target pose are connected, and by comparing a plurality of "arc-line-arc" combinations, the shortest path is selected as the optimal horizontal path; Taking the horizontal flight distance as the reference, the horizontal flight distance is the optimal horizontal path, and the optimal vertical change profile corresponding to the start point height and climbing angle to the target height and climbing angle and the end point height and climbing angle to the target height and climbing angle is planned; The optimal horizontal path and the optimal vertical change profile are fused to reconstruct a three-dimensional space kinematic trajectory that satisfies all start point and end point state constraints, which is the dynamic gliding flight trajectory. 10.The method of claim 1, wherein, After the current energy acquisition flight strategy task is completely finished, and the front potential wind source area is a false wind source area, the UAV cluster switches back to the cruising waiting state, and the latest wind source information in the dynamic wind field information library at the next moment is re-evaluated to decide the next stage of energy acquisition flight.

Images