Method and system for long-endurance unmanned aerial vehicle surveillance mission planning in complex environment

Abstract

The present application relates to the technical field of path planning, in particular to a long-endurance unmanned aerial vehicle monitoring task planning method and system for complex environment, a dynamic wind field model and a target random motion model are constructed according to the wind field change in the task area and the target random motion characteristics; and a unmanned aerial vehicle motion model and an energy consumption model are constructed by using Dubins model and energy consumption theory; the task requirements and constraint conditions of different task stages are set according to the dynamic wind field model, the target random motion model, the unmanned aerial vehicle motion model and the energy consumption model, and the corresponding task optimization model is established, the sum of the overall performance benefits of the unmanned aerial vehicle cluster motion is taken as the optimization objective function to optimize and solve the task optimization model, and the optimal execution strategy corresponding to the task execution stage is obtained according to the solving result. The present application can solve the problem of long-endurance unmanned aerial vehicle cluster cooperative search monitoring task planning for complex environment, and can improve the unmanned aerial vehicle cluster task execution efficiency under the condition of limited energy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of route planning, in particular to a long-endurance unmanned aerial vehicle monitoring task planning method and system for complex environment. BACKGROUND

[0002] Unmanned aerial vehicles (UAVs) have been increasingly used to extend the capabilities of manned aircraft in modern air warfare. Efficient cooperation among multiple UAVs can effectively improve the task execution efficiency of the UAV swarm. The distribution, autonomy and robustness of the cooperative control of the UAV swarm require the decentralized and self-organizing characteristics of biological groups such as ant colonies, bee colonies, bird colonies and fish colonies. The UAV swarm is an autonomous system inspired by the self-organizing mechanism of biological groups, which greatly improves the execution efficiency through complementary capabilities and cooperative combat. Cooperative control technology of UAV swarm has been widely studied and used in various applications such as swarm clustering, reconnaissance, surveillance, load transportation, search, task allocation and path planning.

[0003] In the research field of multi-UAV cooperative mission planning problem (MCMPP), from the perspective of task type, the path planning research for static tasks is relatively mature. The main classical models include multi-trip salesman model (MTSP), mixed integer linear programming model (MILP) and vehicle scheduling and path planning model VRP. These models are solved by heuristic algorithms such as genetic algorithm, evolutionary algorithm, ant colony algorithm and particle swarm algorithm. In the research of task path optimization for dynamic tasks, the task environment is often dynamic and unknown, and the path planning method based on static model cannot be directly applied. Therefore, before the task path optimization, the complex task environment constraints need to be modeled. In particular, the endurance and performance of the UAV swarm system are fundamentally limited by the on-board energy, which is actually limited due to the size and weight restrictions of the aircraft. In theory, in a surveillance mission, the UAV should remain stationary at the location closest to the target to maintain the best surveillance conditions. However, in terms of propulsion energy consumption, hovering at strictly zero speed is inefficient (for rotary-wing UAVs) and even impossible (for fixed-wing UAVs). Therefore, there is an urgent need for a long-endurance UAV swarm cooperative search and surveillance task planning scheme for complex environment to implement the UAV swarm search and surveillance task in a dynamic task environment and optimize the task execution efficiency of the UAV swarm. SUMMARY

[0004] To this end, the present application provides a long-endurance unmanned aerial vehicle monitoring task planning method and system for complex environment, which solves the problem of long-endurance UAV swarm cooperative search and surveillance task planning for complex environment, and can improve the task execution efficiency of the UAV swarm and the monitoring time of the target under the condition of limited energy.

[0005] According to the design scheme provided by the application, a long-endurance unmanned aerial vehicle monitoring task planning method for complex environment is provided, comprising:

[0006] A dynamic wind field model and a target random motion model are constructed according to the wind field change in the task area and the target random motion characteristics, and a unmanned aerial vehicle motion model and an energy consumption model are constructed by using Dubins model and energy consumption theory;

[0007] According to the task description, the task content is divided into a search phase, a monitoring phase and a return phase, and the task requirements and constraint conditions of different task phases are set according to the dynamic wind field model, the target random motion model, the unmanned aerial vehicle motion model and the energy consumption model;

[0008] According to the task requirements and constraint conditions of different task phases, the corresponding task optimization model is established, and the sum of the overall performance benefits of the unmanned aerial vehicle cluster motion is taken as the optimization objective function to optimize and solve the task optimization model, and the optimal execution strategy corresponding to the task execution phase is obtained according to the solving result.

[0009] As the long-endurance unmanned aerial vehicle monitoring task planning method for complex environment of the application, further, a dynamic wind field model is constructed according to the wind field change in the task area, comprising:

[0010] Firstly, the wind field in the task area is divided into an average wind field and a turbulent wind field according to the engineering model;

[0011] Then, a dynamic wind field model is constructed according to the Delton turbulent model, and the wind field distribution in the task area and the component of the wind field in the coordinate axis direction are obtained according to the dynamic wind field model.

[0012] As the long-endurance unmanned aerial vehicle monitoring task planning method for complex environment of the application, further, a target random motion model is constructed according to the target random motion characteristics, comprising:

[0013] Firstly, a basic motion model of the dynamic target is constructed according to the dynamic target motion distance and motion direction and by using the polar coordinate system parameter equation;

[0014] Then, the artificial potential field attraction force and the motion driving force of the dynamic target in the motion process in the task area are obtained according to the artificial potential field algorithm, and the control influence resultant force of the dynamic target motion trajectory is obtained by summing the artificial potential field attraction force and the motion driving force;

[0015] Then, referring to the Wiener random process and according to the basic motion model and the control influence resultant force, the random radian value and the motion speed value are set in the dynamic target motion state equation and the random motion model of the dynamic target is constructed.

[0016] As the long endurance unmanned aerial vehicle monitoring task planning method for complex environment of the present application, further, the Dubins model is used to construct the unmanned aerial vehicle motion model, the Dubins model is used to approximate the unmanned aerial vehicle model, and the Dubins shortest path is set in the flight process of the unmanned aerial vehicle, and the displacement distance, yaw angle, heading angle, minimum turning radius and maximum tilt angle in the flight process of the unmanned aerial vehicle are used to construct the unmanned aerial vehicle motion model.

[0017] As the long endurance unmanned aerial vehicle monitoring task planning method for complex environment of the present application, further, the energy consumption theory is used to construct the unmanned aerial vehicle energy consumption model, and the unmanned aerial vehicle energy consumption model is constructed according to the unmanned aerial vehicle energy consumption influencing factors and the dynamic wind field model in the task area, wherein the unmanned aerial vehicle energy consumption influencing factors include flight time and time consumed in turning at the minimum turning radius in the flight process.

[0018] As the long endurance unmanned aerial vehicle monitoring task planning method for complex environment of the present application, further, the task requirements and constraint conditions of different task stages are set according to the dynamic wind field model, the target random motion model, the unmanned aerial vehicle motion model and the energy consumption model, including: in the search stage, the dynamic target is set to move in the task area, the unmanned aerial vehicle cluster goes to the task area to perform the search task, and if the dynamic target is within the maximum detection distance of the unmanned aerial vehicle, the target is set to be searched and found, and the target position information is shared to the unmanned aerial vehicle cluster; in the monitoring stage, the distance between the unmanned aerial vehicle and the dynamic target is set to be less than the maximum detection distance of the unmanned aerial vehicle as the constraint condition of the monitoring task, and the flight speed of the unmanned aerial vehicle is adjusted according to the dynamic target motion speed, and the distance between the unmanned aerial vehicle and the no-fly zone is set to be greater than the preset value in the monitoring process, and the unmanned aerial vehicle has reserved energy for return; in the return stage, the constraint condition of the remaining energy of the unmanned aerial vehicle is set according to the distance between the unmanned aerial vehicle and the corresponding airport, the unit displacement energy consumption in the flight process of the unmanned aerial vehicle, the displacement distance and the reserved amount of the preset return energy of the unmanned aerial vehicle.

[0019] As the long endurance unmanned aerial vehicle monitoring task planning method for complex environment of the present application, further, the task optimization model is established according to the task requirements and constraint conditions of different task stages, including: in the search stage and the return stage, the unmanned aerial vehicle is set to fly to the given task point at the fastest speed; in the monitoring stage, the unmanned aerial vehicle is set to fly at the minimum energy consumption under the condition of meeting the monitoring task constraint condition.

[0020] As the long endurance unmanned aerial vehicle monitoring task planning method for complex environment of the present application, further, in the task optimization model of the search stage and the return stage, the attraction of the given task point to the unmanned aerial vehicle is established according to the artificial potential field algorithm, the supplementary angle between the motion direction vector of the unmanned aerial vehicle and the attraction is taken as the benefit evaluation sub-function, and the sum of the overall performance benefits of the unmanned aerial vehicle cluster motion Q times is taken as the optimization objective function in the particle swarm algorithm.

[0021] As the long-endurance unmanned aerial vehicle monitoring task planning method for complex environment of the application, further, in the task optimization model of the monitoring stage, the energy consumption benefit evaluation sub-function is constructed according to the energy consumption of the uniform speed flight of the unmanned aerial vehicle in the Dubins curve model and the energy consumption in the existing wind field, and the sum of the overall performance benefits of the unmanned aerial vehicle cluster motion Q times is taken as the optimization objective function in the particle swarm algorithm.

[0022] Further, the application also provides a long-endurance unmanned aerial vehicle monitoring task planning system for complex environment, comprising: a model construction module, a task division module and an optimization solution module, wherein,

[0023] The model construction module is used for constructing a dynamic wind field model and a target random motion model according to the wind field change and the target random motion characteristics in the task area, and constructing an unmanned aerial vehicle motion model and an energy consumption model by using a Dubins model and an energy consumption theory;

[0024] The task division module is used for dividing the task content into a search stage, a monitoring stage and a return stage according to the task description, and setting the task requirements and constraint conditions of different task stages according to the dynamic wind field model, the target random motion model, the unmanned aerial vehicle motion model and the energy consumption model;

[0025] The optimization solution module is used for establishing the corresponding task optimization model according to the task requirements and constraint conditions of different task stages, and optimizing and solving the task optimization model by taking the sum of the overall performance benefits of the unmanned aerial vehicle cluster motion as the optimization objective function, and obtaining the optimal execution strategy corresponding to the task execution stage according to the solution result.

[0026] The application has the following beneficial effects:

[0027] The application can effectively improve the energy utilization efficiency of the unmanned aerial vehicle and the task execution efficiency of the unmanned aerial vehicle cluster by modeling the task description, establishing the dynamic wind field model, the unmanned aerial vehicle motion model and the energy consumption model, and optimizing the flight speed and flight trajectory of the unmanned aerial vehicle according to the task requirements of different stages by using the RHC-PSO rolling horizon control-particle swarm optimization algorithm, and can as much as possible improve the monitoring time of the target of the unmanned aerial vehicle cluster under the condition of limited energy. Further, the simulation results show that the application can optimize the speed adjustment of the unmanned aerial vehicle, effectively improve the endurance and the monitoring task time of the unmanned aerial vehicle, can improve the monitoring task time by 46.9% and the energy utilization efficiency of the unmanned aerial vehicle by 40.2% compared with the task planning method without adjusting the speed of the unmanned aerial vehicle, and has good application prospect in the applications such as group clustering, reconnaissance, monitoring, load transportation, search, task allocation and path planning. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 A schematic diagram of a long-endurance unmanned aerial vehicle surveillance task planning process in the embodiment facing complex environment is shown in the figure;

[0029] Figure 2 A schematic diagram of a search surveillance task decision-making process in the embodiment is shown in the figure;

[0030] Figure 3 A schematic diagram of a particle swarm structure in the RHC-PSO algorithm in the embodiment is shown in the figure;

[0031] Figure 4 A schematic diagram of the RHC-PSO algorithm process in the embodiment is shown in the figure;

[0032] Figure 5 A schematic diagram of a wind field intensity image established based on the Dryden wind field model in the embodiment is shown in the figure;

[0033] Figure 6 A schematic diagram of a minimum distance change curve between the unmanned aerial vehicle cluster and the no-fly zone in the embodiment is shown in the figure;

[0034] Figure 7 A schematic diagram of the unmanned aerial vehicle cluster motion trajectory generated in the embodiment is shown in the figure;

[0035] Figure 8 A schematic diagram of the unmanned aerial vehicle cluster motion trajectory generated by other existing methods in the embodiment is shown in the figure;

[0036] Figure 9 A schematic diagram of a task execution condition comparison in the comparative scheme in the embodiment is shown in the figure;

[0037] Figure 10 A schematic diagram of a comparison of the unmanned aerial vehicle cluster flight energy consumption curves in the comparative scheme in the embodiment is shown in the figure;

[0038] Figure 11 A schematic diagram of a comparison of the unmanned aerial vehicle cluster surveillance durations in 30 Monte Carlo simulations in the embodiment is shown in the figure;

[0039] Figure 12 A schematic diagram of a comparison of the unmanned aerial vehicle cluster energy utilization efficiencies in 30 Monte Carlo simulations in the embodiment is shown in the figure. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical scheme and advantages of the present application more clear, specific, the present application is further described in detail below with reference to the drawings and technical scheme.

[0041] The embodiment of the present application is related to complex environment, unmanned aerial vehicle cluster task planning and unmanned aerial vehicle energy consumption model, and fully considers the influence of complex task environment on task planning, as shown in the figure, Figure 1 A complex environment-oriented long-endurance unmanned aerial vehicle surveillance task planning method is provided, which comprises:

[0042] S101. Based on the wind field changes and target random motion characteristics within the mission area, construct a dynamic wind field model and a target random motion model; and use the Dubins model and energy consumption theory to construct a UAV motion model and an energy consumption model.

[0043] Conventional flight simulations are based on the calm atmosphere assumption, which assumes zero wind speed and no wind interference. However, in actual flight, wind interference poses a significant threat to flight safety, thus requiring an explicit description of wind disturbances during simulation. To describe the randomness and intermittent nature of wind, based on an engineering model, the wind field within the mission area can be approximated as consisting of two parts: mean wind and turbulent wind field.

[0044] Therefore, further in this embodiment, constructing a dynamic wind field model based on wind field changes within the task area includes:

[0045] First, based on the engineering model, the wind field within the task area is divided into the mean wind field and the turbulent wind field;

[0046] Then, a dynamic wind field model is constructed based on the Dryden turbulence model, and the wind field distribution and wind field components in the coordinate axis direction within the task area are obtained based on this dynamic wind field model.

[0047] Mean wind is the baseline value of wind speed, which varies with time and space. A more precise definition is the average wind speed over a specific period. In engineering research, the variation of mean wind over time can be disregarded, and the statistical value of mean wind speed can be used as an approximation. Turbulent wind is a continuous random fluctuation superimposed on the mean wind, equivalent to a random quantity constantly fluctuating around an average value. The wind field model shown in formula (1) can be obtained from the Dryden turbulence model:

[0048]

[0049]

[0050]

[0051]

[0052] a i,j ,b i,j ,c i,j ,α i,j ,β i,j ∈[-1,1]

[0053] in, Represents the two-dimensional vector matrix of mean wind. Represents a two-dimensional vector matrix of turbulent wind. Let a represent the average wind vector at position (i,j). i,j,b i,j ,c i,j α is a constant. i,j ,β i,j The number is random. The wind field distribution map within the task area and the wind field components in the X and Y axes can be obtained through formula (1).

[0054] Furthermore, in this embodiment, constructing a target random motion model based on the target's random motion characteristics includes:

[0055] First, a basic motion model of the dynamic target is constructed based on the dynamic target's motion distance and direction, using the parametric equations of the polar coordinate system;

[0056] Next, the artificial potential field algorithm is used to obtain the attractive force of the artificial potential field in the task area and the driving force of the motion in the direction of motion of the dynamic target during the motion process. The resultant force of the control influence on the motion trajectory of the dynamic target is obtained by summing the attractive force of the artificial potential field and the driving force of motion.

[0057] Then, drawing on Wiener's stochastic process and based on the basic motion model and the resultant force of control influence, random radian values ​​and motion velocity values ​​are set in the motion state equation of the dynamic target, and a stochastic motion model of the dynamic target is constructed.

[0058] Assume that the dynamic target B is an electronic reconnaissance vehicle, patrolling and moving randomly near location C. A motion model of the electronic reconnaissance vehicle based on the artificial potential field algorithm and Wiener stochastic process is established, which can be designed as follows:

[0059] First, a basic motion model of the electronic reconnaissance vehicle is established. Assuming the vehicle's step size is τ, the motion state of the electronic reconnaissance vehicle is described using polar coordinate system parametric equations. The motion state of the electronic reconnaissance vehicle during its k-th movement is {path}. B (k),θ B (k)} represents the distance and direction of movement, respectively.

[0060] Then, to ensure that the electronic reconnaissance vehicle's trajectory is distributed around position C, according to the artificial potential field algorithm, the coordinates of position C are assumed to be C(x). c ,y c If the electronic reconnaissance vehicle is in B(x), then the electronic reconnaissance vehicle is in B(x). B (k),y B (k) is affected by C(x) c ,y c Artificial potential field attraction Meanwhile, the electronic reconnaissance vehicle is subject to θ B Motion driving force in the (k-1) direction The summation of the two vectors in the Cartesian coordinate system is shown in formula (2):

[0061]

[0062] wherein, and are unit vectors, λ represents the weight and direction of the attractive force d CB > d0 represents that the distance between the electronic reconnaissance vehicle and the position C is large, at this time the electronic reconnaissance vehicle is subjected to the attractive force from the point C(x c , y c ), and vice versa, the electronic reconnaissance vehicle is subjected to the repulsive force from the point C(x c , y c ), so that the motion trajectory of the electronic reconnaissance vehicle is around the circle with C(x c , y c ) as the center and d0 as the radius.

[0063] The Cartesian coordinate equation of the resultant force can be derived from formula (2), and the polar coordinate equation of the resultant force is converted to generate formula (3) as shown below:

[0064]

[0065]

[0066] Finally, in order to make the motion state of the electronic reconnaissance vehicle have randomness while meeting the maneuvering characteristics, referring to the Wiener random process, a random radian value and a motion speed value v B within a certain threshold are added to the motion state iteration equation. Thus, the trajectory generation formula (4) of the electronic reconnaissance vehicle is as shown below:

[0067]

[0068] Further, the embodiment of the present case adopts the Dubins model to construct the unmanned aerial vehicle motion model, adopts the Dubins model to approximate the unmanned aerial vehicle model, and sets to fly according to the Dubins shortest path in the flight process of the unmanned aerial vehicle, and uses the displacement distance, yaw angle, heading angle, minimum turning radius and maximum tilt angle in the flight process of the unmanned aerial vehicle to construct the unmanned aerial vehicle motion model.

[0069] The energy consumption in the flight process of the unmanned aerial vehicle mainly includes motion energy consumption and communication energy consumption, wherein the energy consumption related to communication is usually much smaller than the motion energy consumption of the unmanned aerial vehicle, and thus is ignored here. It is assumed that the mode of the indicated airspeed of the unmanned aerial vehicle in the task execution process is and the flight height is constant. Considering that the UAV is a fixed-wing UAV, there is a minimum turning radius constraint, a Dubins model is used to approximate the kinematic model thereof, and it is assumed that the UAV flies according to a Dubins shortest path during flight, i.e., the path between any two points is a straight line segment or a shortest path flight composed of a minimum turning radius corresponding circular arc and straight line segment, so as to save flight time. Based on this, the kinematic model of the UAV can be described as shown in formula (5):

[0070]

[0071] wherein ε(k) represents the displacement distance of the UAV at time k, represents the yaw angle of the UAV at time k, θ(k) represents the heading angle of the UAV at time k, and r(k) represents the minimum turning radius of the UAV at time k, and κ represents the maximum roll angle of the UAV.

[0072] Further, the energy consumption model of the UAV is constructed by using the energy consumption theory, and the energy consumption model of the UAV is constructed according to the energy consumption influencing factors of the UAV and the dynamic wind field model in the task area, wherein the energy consumption influencing factors of the UAV include flight time and time consumed for turning at the minimum turning radius during flight.

[0073] When the fixed-wing UAV flies at a constant speed, the energy consumption of the UAV is only related to the flight speed and acceleration, i.e.,

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] wherein a(k) represents the module of the acceleration of the UAV at time k. c1 and c2 are constant parameters, which are affected by factors such as the weight of the UAV, the wing area, and the air density. V dm represents the minimum flight speed of the UAV. The values of other parameters involved in the formula and the descriptions thereof are shown in Table 2.

[0080] Based on the above analysis, it can be seen that when the flight speed of the unmanned aerial vehicle is constant, the energy consumption of the unmanned aerial vehicle is mainly affected by two aspects. One is the total flight time, and the other is the time consumed in turning with the smallest turning radius during flight. Considering the dynamic wind field distributed in the task area, if the thrust provided by the wind field is correctly utilized, the energy of the unmanned aerial vehicle can be saved, and the flight time is prolonged. On the contrary, the energy consumption of the unmanned aerial vehicle will increase. According to aerodynamics, when there is a wind field that is not 0, the energy consumption of the unmanned aerial vehicle should use the module of the relative air speed instead of the module of the air speed to indicate the air speed. The relationship between the two is shown in formula (7): The relationship between the two is shown in formula (7):

[0081]

[0082] S102, according to the task description, the task content is divided into a search phase, a monitoring phase and a return phase, and the task requirements and constraint conditions of different task phases are set according to the dynamic wind field model, the target random motion model, the unmanned aerial vehicle motion model and the energy consumption model.

[0083] The purpose of the search and monitoring task of the unmanned aerial vehicle cluster is to search and monitor unknown dynamic targets under given constraint conditions, and then to improve the monitoring time of the target as much as possible. According to the task description, the task can be divided into three phases, namely the search phase, the monitoring phase and the return phase. In the embodiment, the task requirements and constraint conditions of different task phases are set according to the dynamic wind field model, the target random motion model, the unmanned aerial vehicle motion model and the energy consumption model, including: in the search phase, the dynamic target is set to move in the task area, the unmanned aerial vehicle cluster goes to the task area to perform the search task, and if the dynamic target is within the maximum detection distance of the unmanned aerial vehicle, the target is set to be searched and found, and the target position information is shared to the unmanned aerial vehicle cluster; in the monitoring phase, the distance between the unmanned aerial vehicle and the dynamic target is set to be less than the maximum detection distance of the unmanned aerial vehicle as the constraint condition of the monitoring task, and the flight speed of the unmanned aerial vehicle is adjusted according to the motion speed of the dynamic target, and the distance between the unmanned aerial vehicle and the no-fly zone is set to be greater than a preset value during the monitoring process, and the unmanned aerial vehicle reserves return energy; in the return phase, the constraint condition of the remaining energy of the unmanned aerial vehicle is set according to the distance between the unmanned aerial vehicle and the corresponding airport, the unit displacement energy consumption of the unmanned aerial vehicle during flight, the displacement distance and the reserved amount of the adjusted unmanned aerial vehicle return energy.

[0084] In the search phase, the dynamic target B is known to move around position C. Therefore, the unmanned aerial vehicle cluster should go to position C to perform the search task as soon as possible. The maximum detection distance of the unmanned aerial vehicle is R, and the target is considered to be searched and found if it appears within the maximum detection distance of the unmanned aerial vehicle. When the target is searched and found by any unmanned aerial vehicle, the position information of the target is shared in the unmanned aerial vehicle cluster. The unmanned aerial vehicle that meets the monitoring task constraint C m will turn into the monitoring phase. The unmanned aerial vehicle that does not meet C​m The drones then move towards the target location based on the target location information shared within the cluster, until the surveillance task constraint C is met. m After that, it entered the monitoring phase.

[0085] C m :d u,C ≤R,u={1,2} (8)

[0086] Where, d u,C This represents the distance between the drone and the target C. u = {1, 2} represents the drone number.

[0087] During the surveillance phase, the drone swarm should adjust its flight speed appropriately based on the target's movement speed to reduce energy consumption and maximize surveillance duration. Simultaneously, constraint C should be considered. m Irregular no-fly zone constraints C Obs and return-to-base energy constraints C E The distance between the drone and the no-fly zone must be greater than 1, and the drone must have sufficient energy reserves for its return flight. Constraint C Obs and C E As shown in formula (9):

[0088]

[0089] Where, d u,Obs Let d represent the distance between the drone and the target C. Let E(k) represent the drone's remaining energy at time k. u,ap This indicates the distance between the drone and its corresponding airport. E mean (k) and d u These represent the energy consumption per unit displacement and the displacement distance of the UAV from the start of the mission to time k, respectively. λ is a constant parameter used to adjust the reserve of energy for the UAV's return journey.

[0090] Because wind fields are dynamic, it is impossible to accurately predict the exact amount of energy required for return. Therefore, a method of estimating energy consumption per unit displacement and adding a reserve is used to ensure that the drone can return with sufficient energy. When the drone cannot meet constraint C... E At that time, the drone entered the return-to-home phase.

[0091] During the return phase, the drone increases its flight speed to return as quickly as possible. During this phase, the drone must meet constraint C. Obs See also Figure 2 As shown, the mission ends when all drones in the fleet return to their respective airports.

[0092] S103, a task optimization model is established according to the task requirements and constraint conditions of different task stages, and the sum of the overall performance benefits of the unmanned aerial vehicle cluster movement is taken as an optimization objective function to optimize the task optimization model, and the optimal execution strategy corresponding to the task execution stage is obtained according to the solving result.

[0093] The task requirements and constraints of three different stages in the search monitoring task execution process are analyzed, the corresponding task optimization model is established, and the optimal solution is generated using an optimization algorithm. In the search stage and the return stage, the unmanned aerial vehicle is set to fly to the predetermined task point at the fastest speed; in the monitoring stage, the unmanned aerial vehicle is set to fly at the minimum energy consumption under the condition of meeting the monitoring task constraints.

[0094] The RHC (Rolling Horizon Control)-PSO (Particle Swarm Optimization) algorithm is used to optimize the task decision. On the basis of the traditional particle swarm optimization algorithm, in order to improve the global optimization effect and obstacle avoidance effect of the unmanned aerial vehicle, the rolling horizon decision method in the model predictive control theory is used. First, the corresponding optimization function is established according to the optimization objective of different task stages. Then, the sum of the overall performance benefits of the unmanned aerial vehicle cluster movement Q times is taken as the optimization objective function to optimize. The particle swarm structure and algorithm flow in the RHC-PSO algorithm are shown in Figure 3 and Figure 4 In the task optimization model of the search stage and the return stage, the attractive force of the predetermined task point on the unmanned aerial vehicle is established according to the artificial potential field algorithm, the supplementary angle between the unmanned aerial vehicle movement direction vector and the attractive force is taken as the benefit evaluation sub-function, in the task optimization model of the monitoring stage, the energy consumption benefit evaluation sub-function is constructed according to the energy consumption of the unmanned aerial vehicle uniform flight in the Dubins curve model and the energy consumption in the presence of wind field, and the sum of the overall performance benefits of the unmanned aerial vehicle cluster movement Q times is taken as the optimization objective function in the particle swarm algorithm.

[0095] Figure 4 In the equation, p bestm and g best are the historical optimal solution of the particle swarm and the population optimal solution of the particle swarm respectively, psv m is the particle swarm movement speed, ω is the inertia factor, c1 and c2 are the learning factors, r1, r2 ∈ [0, 1] are random numbers, L best (k+1) is the particle swarm optimal solution at time k. κ ∈ [1, K] is the iteration number of the particle swarm, J(k) Q is the Q-time rolling optimization objective function corresponding to the task stage.

[0096] In the search phase and the return phase, the UAV has a clear task goal: flying to the given task point P as soon as possible. In the search phase, the UAV needs to reach the position C as soon as possible for searching. In the return phase, the UAV needs to return to the airport as soon as possible. Therefore, according to the artificial potential field algorithm, the attraction of P to the UAV is established Then, the supplementary angle of the angle between the UAV motion direction vector and is taken as the benefit evaluation sub-function, as shown in formula (10):

[0097]

[0098]

[0099] In the monitoring task phase, the task of the UAV is to reduce the flight energy consumption as much as possible under the condition that C m is met. Due to the existence of the wind field, the UAV needs to utilize the same direction wind field as much as possible to reduce the energy consumption. According to the Dubins curve model, the energy consumption benefit evaluation sub-function is obtained from formula (6) and formula (7), as shown in formula (11):

[0100]

[0101]

[0102]

[0103] Further, based on the above method, the embodiment of the present application also provides a long-endurance UAV monitoring task planning system for a complex environment, comprising: a model construction module, a task division module and an optimization solving module, wherein,

[0104] The model construction module is used for constructing a dynamic wind field model and a target random motion model according to the wind field change and the target random motion characteristics in the task area, and constructing a UAV motion model and an energy consumption model by using a Dubins model and an energy consumption theory;

[0105] The task division module is used for dividing the task content into a search phase, a monitoring phase and a return phase according to the task description, and setting the task requirements and constraint conditions of different task phases according to the dynamic wind field model, the target random motion model, the UAV motion model and the energy consumption model;

[0106] The optimization solving module is used for establishing a corresponding task optimization model according to the task requirements and constraint conditions of different task phases, taking the sum of the overall performance benefits of the UAV cluster motion as an optimization objective function to optimize and solve the task optimization model, and obtaining the optimal execution strategy corresponding to the task execution phase according to the solving result.

[0107] To verify the effectiveness of the scheme, the following test data are further explained:

[0108] Task scenario parameter setting: the task area is 200km x 200km, which is discretized into 200 x 200 grids. Dynamic target B moves randomly around a point C with d0 as the radius in the task area. The motion information of dynamic target B is shown in Table 1. All motion data of the target are randomly generated using the Monte Carlo method.

[0109] Table 1 Motion parameters of dynamic target B

[0110]

[0111] It is assumed that the UAV cluster enters the task area at k = 1, and the system iteration step is 60 seconds. The UAV cluster makes a displacement and performs an environment detection once the system iterates. The UAV cluster contains 2 UAVs with the same performance, and the UAV is equipped with multi-functional sensors such as environment detection, target monitoring, distance calculation, and wind field detection. The values of the specific physical performance parameters involved in formula (6) and their descriptions are given in Table 2.

[0112] Table 2 Values of UAV performance parameters and their descriptions

[0113]

[0114]

[0115] The search and surveillance task execution efficiency under the assumed task conditions is affected by multiple modules in the task planning scheme. The task decision algorithm based on RHC-PSO algorithm proposed in the scheme optimizes the flight speed of the UAV in different task stages to achieve the purpose of reducing the flight energy consumption of the UAV and improving the monitoring task duration. In order to study the task execution efficiency and the performance of the optimization algorithm, several simulations are performed. In the simulation, the scheme is compared with the task planning method without adjusting the flight speed, and the comparison is made in terms of the energy utilization efficiency of the UAV and the UAV task execution efficiency index.

[0116] Figure 5 and 6 The definition and modeling of complex task scenarios in the assumed conditions of the scheme are shown. The wind field intensity image based on the Dryden wind field model is shown in Figure 5 , where (a) represents the wind field vector value, and (b) represents the wind field vector modulus. From the figure, it can be seen that the wind field in the task area has dynamic characteristics, which will affect the flight of the UAV. Therefore, the UAV must consider the influence of the wind field on the flight performance of the UAV when deciding the flight path.

[0117] In addition, there are no-fly zones in the task area, and the UAV swarm should avoid the no-fly zones when optimizing the flight trajectory. Figure 6 As shown in FIG. 3, the distance between the UAV swarm and the no-fly zone is always kept above the minimum safety distance.

[0118] However, the dynamic target can move within the no-fly zone, which makes the UAV unable to directly approach the target when performing the search and surveillance task, but instead continuously monitor the target while keeping a certain distance from the target. The UAV swarm flight trajectory generated based on the scheme of the present application is shown in FIG. 4. The UAV swarm flight trajectory generated based on the task planning method without adjusting the flight speed is shown in FIG. 5. Figure 7 Figure 8

[0119] As can be seen from FIG. 6 and FIG. 7, the UAV without adjusting the flight speed can monitor the target, but causes a large amount of flight energy consumption due to the large movement step. Therefore, the monitoring time of the UAV to the target is short. In the UAV swarm flight trajectory generated using the scheme of the present application, the flight speed of the UAV is dynamically adjusted, so that the UAV swarm can more efficiently cooperatively monitor the target. The energy utilization efficiency of the UAV is significantly improved. As shown in FIG. 8, in this simulation, the monitoring time of the UAV swarm with dynamically adjusted flight speed to the target is 858 minutes, while the monitoring time of the other method is 346 minutes. As shown in FIG. 9, the UAV swarm with dynamically adjusted flight speed consumes 97.6% of energy at k = 1006 and successfully returns, and the other method consumes 98.9% of energy at k = 466 and successfully returns. Figure 7 Figure 8 In order to more accurately evaluate the effect of the two methods on improving the endurance of the UAV, 30 Monte Carlo simulations are performed. The two methods are compared from the aspects of task monitoring time and energy utilization efficiency of the UAV. Figure 9 Figure 10 The variation curve of the monitoring task time of the UAV swarm in multiple simulations is shown in FIG. 10. The solid line is the average value of the monitoring task time in multiple simulations. As can be seen from FIG. 10, the average monitoring task time of the UAV swarm with dynamically adjusted flight speed is 868.7 minutes, while the average monitoring task time of the other method is 591.6 minutes. The method of the present application improves by 46.9% compared with the other method.

[0120]

[0121] Figure 11 Figure 11

[0122] Figure 12 ​​​​​​​Fig. 6 shows the energy utilization efficiency of the UAV swarm in 30 simulations. The energy utilization efficiency is the ratio of the total movement distance of the UAV swarm to the total energy of the UAV swarm, in unit of m / KJ. The solid line is the average of the energy utilization efficiency in multiple simulations. It can be seen from Fig. 6 that the average energy utilization efficiency of the UAV swarm with dynamic adjustment of flight speed is 5.51 m / KJ, while that of the other method is 3.93 m / KJ. The method of the present patent improves 40.2% compared with the other method. Figure 11

[0123] Therefore, the multiple simulation data under the multiple constraints including wind field, no-fly zone and limited flight energy show that the method of the present patent has superiority in both energy utilization efficiency and task execution efficiency compared with other methods.

[0124] The relative arrangement, numerical expressions and numerical values of the components and steps set forth in the embodiments are not intended to limit the scope of the present patent, unless otherwise specified.

[0125] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiments, the description is relatively simple because it corresponds to the method disclosed in the embodiments, and the relevant parts can be referred to the description of the method.

[0126] The units and method steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example are generally described in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation does not exceed the scope of the present patent.

[0127] Those skilled in the art can understand that all or part of the steps in the above method can be instructed by a program to complete the relevant hardware, and the program can be stored in a computer readable storage medium, such as a read-only memory, a magnetic disk or an optical disk, etc. Alternatively, all or part of the steps of the above embodiments can also be implemented using one or more integrated circuits, and accordingly, each module / unit in the above embodiments can be implemented in the form of hardware or in the form of a software function module. The present patent is not limited to any specific form of combination of hardware and software.

[0128] ​Finally, it should be noted that the above-described embodiments are merely specific embodiments of the present application, which are used to illustrate the technical solutions of the present application, but not to limit the same. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that any person skilled in the art can still modify or easily think of changes to the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some of the technical features, within the technical scope disclosed by the present application. The modifications, changes or replacements do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for planning long-endurance UAV surveillance missions in complex environments, characterized in that, Include: Based on the wind field changes and target random motion characteristics within the mission area, a dynamic wind field model and a target random motion model are constructed; and a UAV motion model and an energy consumption model are constructed using the Dubins model and energy consumption theory. Based on the task description, the task content is divided into a search phase, a monitoring phase, and a return-to-home phase. Task requirements and constraints for each phase are set according to a dynamic wind field model, a target random motion model, a UAV motion model, and an energy consumption model. Specifically, the task requirements and constraints based on these models include: In the search phase, dynamic targets are set to move within the task area. If a dynamic target is within the maximum detection range of a UAV cluster while performing a search within the task area, the target is considered detected, and its location information is shared with the UAV cluster. In the monitoring phase, the distance between the UAV and the dynamic target is set to be less than the UAV's maximum detection range as a constraint for the monitoring task. The UAV's flight speed is adjusted according to the target's speed, and the distance between the UAV and the no-fly zone is set to be greater than a preset value, with the UAV having reserved return-to-home energy. In the return-to-home phase, constraints on the UAV's remaining return-to-home energy are set based on the distance between the UAV and the corresponding airport, the energy consumption per unit displacement during UAV flight, the displacement distance, and the preset reserved amount of return-to-home energy. Based on the task requirements and constraints of different task stages, corresponding task optimization models are established, and the sum of the overall performance benefits of UAV swarm movement is used as the optimization objective function to optimize and solve the task optimization model. Based on the solution results, the optimal execution strategy corresponding to the task execution stage is obtained.

2. The long-endurance UAV surveillance mission planning method for complex environments according to claim 1, characterized in that, A dynamic wind field model is constructed based on wind field changes within the mission area, including: First, based on the engineering model, the wind field within the task area is divided into the mean wind field and the turbulent wind field; Then, a dynamic wind field model is constructed based on the Dryden turbulence model, and the wind field distribution and wind field components in the coordinate axis direction within the task area are obtained based on this dynamic wind field model.

3. The long-endurance UAV surveillance mission planning method for complex environments according to claim 1, characterized in that, Based on the characteristics of the target's random motion, a target random motion model is constructed, including: First, a basic motion model of the dynamic target is constructed based on the dynamic target's motion distance and direction, using the parametric equations of the polar coordinate system; Next, the artificial potential field algorithm is used to obtain the attractive force of the artificial potential field in the task area and the driving force of the motion in the direction of motion of the dynamic target during the motion process. The resultant force of the control influence on the motion trajectory of the dynamic target is obtained by summing the attractive force of the artificial potential field and the driving force of motion. Then, drawing on Wiener's stochastic process and based on the basic motion model and the resultant force of control influence, random radian values ​​and motion velocity values ​​are set in the motion state equation of the dynamic target, and a stochastic motion model of the dynamic target is constructed.

4. The long-endurance UAV surveillance mission planning method for complex environments according to claim 1, characterized in that, In constructing the UAV motion model using the Dubins model, the Dubins model is used to approximate the UAV motion model, and the UAV is set to fly along the Dubins shortest path during flight. The motion model of the UAV is constructed by using the displacement distance, yaw angle, heading angle, minimum turning radius and maximum tilt angle during the flight of the UAV.

5. The long-endurance UAV surveillance mission planning method for complex environments according to claim 1 or 4, characterized in that, In constructing an UAV energy consumption model using energy consumption theory, the UAV energy consumption model is built based on the factors affecting UAV energy consumption and the dynamic wind field model within the mission area. Among them, the factors affecting UAV energy consumption include flight time and the time consumed during flight when turning with the minimum turning radius.

6. The long-endurance UAV surveillance mission planning method for complex environments according to claim 1, characterized in that, Based on the task requirements and constraints of different task phases, corresponding task optimization models are established, including: in the search and return phases, setting the UAV to fly to the designated task marker at the fastest speed; in the surveillance phase, setting the UAV to fly with the minimum energy consumption while meeting the surveillance task constraints.

7. The long-endurance UAV surveillance mission planning method for complex environments according to claim 6, characterized in that, In the task optimization model for the search and return phases, the attraction of the predetermined task markers to the UAV is established based on the artificial potential field algorithm. The supplementary angle between the UAV's motion direction vector and the attraction is used as the benefit evaluation sub-function. In the particle swarm algorithm, the sum of the overall performance benefits of the UAV swarm movement Q times is used as the optimization objective function.

8. The long-endurance UAV surveillance mission planning method for complex environments according to claim 6, characterized in that, In the task optimization model during the monitoring phase, an energy consumption benefit evaluation subfunction is constructed based on the energy consumption of the UAV during uniform flight and the energy consumption when there is a wind field in the Dubins curve model. In the particle swarm algorithm, the sum of the overall performance benefits of the UAV swarm movement Q times is used as the optimization objective function.

9. A long-endurance unmanned aerial vehicle (UAV) surveillance mission planning system for complex environments, characterized in that: It includes: a model building module, a task partitioning module, and an optimization and solution module, among which, The model building module is used to construct dynamic wind field models and target random motion models based on wind field changes and target random motion characteristics within the mission area; and to construct UAV motion models and energy consumption models using the Dubins model and energy consumption theory. The task partitioning module is used to divide the task content into three phases based on the task description: search, surveillance, and return. It sets the task requirements and constraints for each phase based on a dynamic wind field model, a target random motion model, a UAV motion model, and an energy consumption model. Specifically, the task requirements and constraints based on these models include: In the search phase, dynamic targets are set to be active within the task area. If a dynamic target is within the maximum detection range of a UAV cluster while performing a search within the task area, the target is considered detected, and its location information is shared with the UAV cluster. In the surveillance phase, the distance between the UAV and the dynamic target is set to be less than the UAV's maximum detection range as a constraint. The UAV's flight speed is adjusted based on the target's speed, and the distance between the UAV and the no-fly zone is set to be greater than a preset value, with the UAV having reserved return energy. In the return phase, constraints on the UAV's remaining return energy are set based on the distance between the UAV and the corresponding airport, the energy consumption per unit displacement during UAV flight, the displacement distance, and the preset reserved amount of return energy. The optimization and solution module is used to establish corresponding task optimization models based on the task requirements and constraints of different task stages. It uses the sum of the overall performance benefits of the UAV swarm movement as the optimization objective function to optimize and solve the task optimization model, and obtains the optimal execution strategy corresponding to the task execution stage based on the solution results.

Images