Unmanned aerial vehicle no-miss area coverage task parameter optimization method

Abstract

The application provides a method for optimizing task parameters of unmanned aerial vehicle (UAV) without missing area coverage, which comprehensively considers the problems of discovery and coverage search, comprehensively considers various factors such as environment, load, UAV, task personnel and the like, introduces relatively perfect influence variables, establishes an optimization target model and a restriction condition model; based on an immune algorithm framework, a task parameter optimization algorithm process and steps are designed, and finally an excellent task parameter group is obtained, so that the search efficiency of the UAV is maximized while the high discovery probability is ensured, and the fast coverage search of the target area without missing is realized.

Description

Technical Field

[0001] This invention belongs to the field of UAV search technology, and in particular relates to a method for optimizing the parameters of a UAV's photoelectric payload to cover an area without omissions. Background Technology

[0002] Traditional search theories are largely based on traditional search units, such as fixed-wing aircraft, helicopters, ships, vehicles, and personnel. In recent years, with the maturation and improvement of UAVs and their payloads, they have exhibited advantageous characteristics such as long endurance, long range, no personnel casualties, and diverse payloads. The application of UAVs in civilian or military target search has gradually increased, demonstrating significant potential for search applications. Therefore, UAV-based target search methods have become an important research direction.

[0003] According to search theory, finding a target within a certain area typically requires ensuring a complete and comprehensive search of the target's area, while guaranteeing effective detection and sensing of the target by the search unit's field of view. Area coverage search is a common target search method for UAVs, widely used in scenarios such as maritime rescue, disaster relief, area reconnaissance, and patrol surveillance. In UAV area coverage search, photoelectric detection equipment is the most common imaging device, offering advantages such as high resolution and long detection range. However, its static field of view is usually limited, necessitating online operation by personnel to complete the search process. However, due to complex limitations imposed by environmental factors, detection payload, the UAV itself, and the physiological state of the personnel, determining an appropriate set of mission parameters to efficiently and comprehensively complete the coverage search and find the target with a high probability presents a complex nonlinear optimization problem that is both practical and challenging.

[0004] Existing mission parameter planning methods generally do not consider the impact of detector resolution and environmental visibility on target detection, and focus primarily on region segmentation and trajectory planning methods. Yu Si-nan et al. (Journal of Beijing University of Aeronautics and Astronautics, 2015, 41(1):167-173) studied the region segmentation and trajectory planning methods for reconnaissance with multiple UAVs covering complex areas. Wu Qingpo et al. (Tactical Missile Technology, 2016(1):50-55) established a geometric relationship model between the static detection width of the UAV detector and the flight altitude, pitch angle, and search azimuth angle. However, none of these methods considered the limitations and influencing factors such as target detection, speed-to-altitude ratio, and scan omissions, which differed significantly from the actual situation when the payload was working. Tan et al. (Ship Electronic Engineering, 2019, 39(06):146-150) started from the meaning of scan width and studied the relationship between indicators such as detection probability and coverage. By analyzing the working mode of the detection equipment and the pixel limitation principle of target recognition, they obtained the scan width calculation method under various detection modes, but did not consider the impact of target detection, missed scans, and visual fatigue of mission personnel. Overall, the existing methods are too simplistic in modeling the scanning range of UAV optoelectronic payloads, and do not consider the problems of missed scans and target detection; the physiological effects of mission personnel are also not considered when planning mission parameters. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for optimizing parameters for unmanned aerial vehicle (UAV) coverage missions with no omissions, comprising the following steps:

[0006] S1. Establish a coverage width model;

[0007] S2. Establish a constraint model;

[0008] S3. Search parameter planning based on immune algorithm.

[0009] Specifically, the steps to establish the coverage width model are as follows:

[0010] S11. Determine the scanning mode of the optoelectronic device;

[0011] S12. Establish a static field-of-view model for optoelectronic equipment;

[0012] S13. Establish a no-omission coverage width model.

[0013] Furthermore, the static field-of-view model of the optoelectronic equipment is established as follows:

[0014] Under the current weather conditions, for the maximum recognition distance of the target to be searched, the position of the optoelectronic device is O1, its projection position on the ground is O, the flight altitude is h, OX is the projection direction of the central axis of the field of view on the ground, OY is the vertical upward direction, and the right-hand rule determines the OZ axis; ε is the size of the field of view angle, μ1 is the depression angle of the inner edge of the field of view, and the area formed by ABCD is the coverage area of ​​the static field of view of the optoelectronic device on the ground, with the lengths of the field of view edges being l. AB l AD =l BC l CD As shown in the following formula:

[0015]

[0016] Furthermore, the specific details of establishing the no-omission coverage width model are as follows:

[0017] Establish a two-dimensional coordinate system horizontal to the ground, with O1 as the origin, the flight direction as the Y-axis, the X-axis perpendicular to the Y-axis, and rightward as positive.

[0018] d1 = 2|x1|

[0019]

[0020] Where R1 and R2 are the inner and outer radii of the static coverage area, V is the flight speed, and ω is the angular velocity of the optical axis of the optoelectronic device in the horizontal plane; when the optoelectronic device moves radially from position O1 to position O2 along the Y-axis, the scanning optical axis rotates 2θ1 from the leftmost position to the rightmost position. When the optoelectronic device moves from position O2 to position O3, the scanning optical axis will rotate 2θ1 from the rightmost position to the original position. This process is one cycle, and the time taken for one scanning cycle is T; within one scanning cycle, the scanning of two arc regions is completed. θ1 is the dynamic search azimuth angle, P1 is the innermost point of the "missed area", denoted by (x1, y1); d1 is the containment width for a search without omissions.

[0021] Furthermore, the specific steps for establishing a constraint model are as follows:

[0022] S21. Determine the speed-to-height ratio limiting conditions;

[0023] S22. Determine the search restrictions for no omissions;

[0024] S23. Determine the limitations on the field of view depression angle, flight altitude, cruise speed, and the angular velocity of the optical axis of the optoelectronic equipment in the horizontal plane.

[0025] Specifically, the speed-to-height ratio limit conditions for S21 are as follows:

[0026]

[0027] The angle of deviation from the flight course is θ′, and γ is the maximum speed-to-height ratio;

[0028] Specifically, the no-omission search constraint in S22 is as follows: the region formed by A1B1C1D1 is the area covered by the static field of view on the ground when the photoelectric device is located at O1; the region formed by A2B2C2D2 is the area covered by the static field of view on the ground when the photoelectric device is located at O2; and the region formed by A3B3C3D3 is the area covered by the static field of view on the ground when the photoelectric device is located at O3. For any x in the interval [-R1sinθ1, R1sin(θ1-ε)], the following holds:

[0029] y c2 (x)≤y B1 (x)

[0030] The functional expression for trajectory B1B2 is:

[0031]

[0032] The functional expression for trajectory C2C3 is:

[0033]

[0034] Furthermore, the specific steps for planning the search parameters based on the immune algorithm in S3 are as follows:

[0035] S31. Establish an optimization index model;

[0036] S32. Optimize the target model based on the immune algorithm.

[0037] Specifically, the optimized indicator model is as follows:

[0038]

[0039] max(J)

[0040] T = 4(θ1 - ε / 2) / ω

[0041]

[0042] R1 = h·tan(arctan(R2 / h)-ε)

[0043] μ2 = arctan(R2 / h)

[0044] μ1=μ2-ε

[0045]

[0046] d1 = 2|x1|

[0047]

[0048]

[0049] ST

[0050] θ1≤90°

[0051] V2≥V≥V1

[0052] h≥h s

[0053]

[0054] μ min ≤μ1

[0055] ω≤ω max

[0056]

[0057] y c2 (x)≤y B1 (x), x∈[-R1sinθ1,R1sin(θ1-ε)]

[0058] Among them, K η The weighting coefficients for coverage efficiency η, K ω The weighting coefficient for the angular velocity ω of the optoelectronic device. The weighting coefficient for the optimal cruising speed V0 The weighting coefficients for the optimal cruising altitude h0 are: |Δh0| and |ΔV0|, which are the deviations between the mission plan's flight altitude h and flight speed V and their optimal values, respectively; and J is the comprehensive optimization objective function, μ2 is the depression angle of the outer edge, and μ min For the lowest top-down view, h max For the highest safe flight altitude, h s For the minimum safe flight altitude, ω max This represents the maximum angular velocity of the optical axis of the optoelectronic device's field of view in the horizontal plane.

[0059] Specifically, the steps of the immune algorithm in S32 are as follows:

[0060] The velocity V, flight altitude h, azimuth angle θ1, and scanning angular velocity ω are combined as a set of parameters to be optimized; each set of parameters is considered an antibody in the algorithm population, with a population size of 50; the objective function value J generated based on the set of task parameters is used as the fitness f of the antibody; the clonal selection probability P... s =0.4, mutation probability P m=0.8, update probability P u =0.2;

[0061] Step 1: Initialize algorithm-related parameters, initialize search area coordinates, initialize UAV platform performance parameters, initialize optoelectronic device performance parameters, and initialize... and;

[0062] Step 2: Based on the threshold range of each parameter, randomly generate 50 sets of task parameter groups to form the initial antibody population. Each set of parameters contains 4 parameters: [V,h,θ1,ω].

[0063] Step 3: Calculate the objective function value J for each antibody in the population as the fitness f of that antibody;

[0064] Step 4: Calculate whether each antibody in the population satisfies the velocity-to-high ratio restriction and the missed scan restriction. If any one of the conditions is not met, the fitness of the antibody is set to 0. If both restrictions are met, the fitness value of the antibody remains unchanged.

[0065] Step 5: Determine if the optimal antibody meets the task requirements. If it does, the algorithm ends; otherwise, proceed to the next step.

[0066] Step 6: Calculate the vector moment concentration of each antibody in the population based on fitness, and calculate the selection probability of that antibody based on the vector moment concentration;

[0067] Step 7: Select antibodies and perform clonal amplification based on the concentration regulation mechanism of the immune algorithm, with a clonal selection probability P. s =0.4;

[0068] Step 8: Perform a mutation operation on each antibody in the cloned and amplified population, with a mutation probability P. m =0.8, every parameter of the antibody obtained after mutation must meet the threshold range limit of that parameter; if it does not meet the limit, it must be mutated again.

[0069] Step 9: Determine if the number of antibodies in the current population has reached 50. If not, randomly generate antibodies to replenish the population to 50.

[0070] Step 10: Determine whether the set conditions for the algorithm to run have been met. If they have, the algorithm ends; otherwise, use the new population obtained as the initial population and go to step 3 to start a new round of calculation.

[0071] The beneficial effects of this invention are that it incorporates the main influencing factors of optoelectronic payload, airframe platform, environment, and mission personnel. A static field-of-view model is established for the optoelectronic equipment, determining the target detection distance threshold and static field-of-view parameters based on target characteristics, environmental factors, and optoelectronic payload resolution. An effective model is established for the coverage area of ​​the UAV's optoelectronic equipment under typical operating modes, constructing a method for determining the coverage width of the UAV and its optoelectronic equipment under typical modes without omissions. A constraint model for omission-free coverage and a speed-to-altitude ratio constraint model are established. An optimization index for search efficiency is proposed, and a comprehensive optimization index is established by combining several commonly used indicators. Based on evolutionary calculation methods and the previously established model, optimization calculations are performed, resulting in mission parameters that meet all constraints while achieving higher search efficiency and preventing omissions. Attached Figure Description

[0072] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0073] Figure 1 This is a schematic diagram of a sector scan.

[0074] Figure 2 This is a schematic diagram of the static field-of-view model of an optoelectronic device.

[0075] Figure 3 This is a schematic diagram of the coverage width model without omissions.

[0076] Figure 4 This is a schematic diagram of the speed-to-height ratio limiting model.

[0077] Figure 5 This is a diagram illustrating the coverage of omissions.

[0078] Figure 6 This is the flowchart of the immune algorithm.

[0079] Figure 7 It is a simulation of the coverage process and coverage area under 6 sets of empirical parameters, among which Figure 7 (a) is a simulation diagram of the coverage process and coverage area when V = 80 m / s, h = 3500 m, ω = 2° / s, and θ1 = 75°. Figure 7 (b) is a simulation diagram of the coverage process and coverage area with V = 70 m / s, h = 2500 m, ω = 3° / s, and θ1 = 60°. Figure 7 (c) is a simulation diagram of the coverage process and coverage area with V = 60 m / s, h = 2000 m, ω = 3° / s, and θ1 = 80°. Figure 7(d) is a simulation diagram of the coverage process and coverage area with V = 83 m / s, h = 3000 m, ω = 4° / s, and θ1 = 80°. Figure 7 (e) is a simulation diagram of the coverage process and coverage area with V = 83 m / s, h = 1500 m, ω = 2.5° / s, and θ1 = 85°. Figure 7 (f) is a simulation diagram of the coverage process and coverage area with V = 83 m / s, h = 5000 m, ω = 2.5° / s, and θ1 = 85°.

[0080] Figure 8 This describes the variation of J with V and h when the azimuth angle θ1 = 85° and the scanning search angular velocity ω = 3° / s.

[0081] Figure 9 This refers to the variation of J with azimuth angle θ1 and scanning search angular velocity ω when V = 80 m / s and h = 3500 m.

[0082] Figure 10 This refers to the variation of J with h and θ1 when V = 80 m / s and ω = 3° / s.

[0083] Figure 11 This refers to the variation of J with V and θ1 when h = 3500m and ω = 3° / s.

[0084] Figure 12 It is the curve of the change of the comprehensive index J, which represents the optimization process.

[0085] Figure 13 This is the curve showing the change in coverage width d1 of the representative optimization process.

[0086] Figure 14 It is a speed change curve of a representative optimization process.

[0087] Figure 15 It is a curve showing the height variation of a representative optimization process.

[0088] Figure 16 It is a simulation of the scanning trajectory of the representative optimization results.

[0089] Figure 17 This is the curve showing the change in scanning efficiency under the final parameters. Detailed Implementation

[0090] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0091] 1. Establish a coverage width model

[0092] Complete and comprehensive coverage of the target area is a prerequisite for searching for the target. Therefore, reasonably determining the coverage width during the search process is of great significance for ensuring the success rate of the search and improving the search efficiency.

[0093] 1.1 Determine the scanning method of the optoelectronic equipment

[0094] UAV optoelectronic payloads have various scanning modes, such as side scanning, cone scanning, and sector scanning. This paper selects the most commonly used and complex sector scanning mode as a typical model; other modes can be considered as special cases of sector scanning. Figure 1 As shown in the figure, assuming the UAV remains in level flight, in the sector scanning mode, the pitch angle μ′ of the optoelectronic device's field of view remains unchanged, and the azimuth angle of the field of view is uniformly scanned back and forth within an angle range [-θ, θ] on the horizontal plane.

[0095] 1.2 Establishing a static field-of-view model for optoelectronic equipment

[0096] In sector scanning mode, the field of view of the optoelectronic device is divided into dynamic field of view and static field of view. The static field of view is the inherent field of view angle of the optoelectronic device, while the dynamic field of view is the range that can be detected when the static field of view is scanning, which is related to the static field of view itself and the azimuth angle of the scan.

[0097] like Figure 2 As shown, let the position of the optoelectronic device be O1, its projection position on the ground be O, the flight altitude be h, OX be the projection direction of the central axis of the field of view on the ground, OY be the vertical upward direction, and the right-hand rule determine the OZ axis.

[0098] Let ε be the field of view angle, and μ1 and μ2 be the depression angles at the inner and outer edges of the field of view, respectively. Then we have:

[0099] μ2=μ1+ε

[0100] If we define μ as the depression angle of the center line of the field of view of the optoelectronic device, then we have:

[0101] μ=(μ1+μ2) / 2

[0102] Under current weather conditions, the maximum identification distance for the target needs to be determined in advance based on weather conditions, target characteristics, the identification range of the optoelectronic equipment, and the experience of the mission personnel. To ensure reliable identification of targets within the search area, there are...

[0103]

[0104] when At this time, the maximum static coverage area can be obtained while ensuring reliable target identification, thus improving search efficiency.

[0105]

[0106] The area formed by ABCD is the coverage area of ​​the static field of view of the optoelectronic device on the ground. It can be seen that the field of view is not a rectangle but a trapezoid; therefore, it is necessary to determine the length of each side to determine the area of ​​the search region.

[0107] The lengths of the sides are respectively represented by l AB l CD l AD It means, l AD =l BC .

[0108] In triangle O1AB, we can obtain

[0109]

[0110] In triangle O1OA, we can obtain

[0111]

[0112] And μ1 = μ2 - ε; in triangle O1OD, we can obtain

[0113] l OD =h·tanμ1

[0114] l O1D =h / cosμ1

[0115] We can obtain edge l CD l AD :

[0116]

[0117] l AD =l OA -l OD

[0118] 1.3 Modeling the width of coverage without omissions

[0119] like Figure 2 As shown, l is defined OA l OD Let R2 be the outer radius and R1 be the inner radius of the static coverage area, then we have

[0120]

[0121] R1 = h·tan(arctan(R2 / h)-ε)

[0122] Let V be the flight speed, and assume that the angular velocity of the optical axis of the optoelectronic device's field of view in the horizontal plane is equal to ω. When the optoelectronic device rotates and scans in the horizontal plane, the depression angle μ of the optical axis of the field of view remains constant.

[0123] like Figure 3 As shown, if we define a scan cycle as the process of the search optical axis rotating from the leftmost end (position O1) to the rightmost end (position O2) and then returning to the leftmost end (position O3), then the time taken for one scan cycle is T = 4(θ1-ε / 2) / ω.

[0124] Within one scan cycle, the drone flies from position O1 to position O3.

[0125] Within one scan cycle, the photoelectric device completes the scan of two arc-shaped regions, the clockwise scanned region shown within the red line, and the counterclockwise scanned region shown within the blue line. Although the search achievable width is...

[0126] d=2R2sinθ1

[0127] However, since the drone is in continuous flight, there is a "missed area" in the direction of the maximum value of the dynamic search azimuth angle between the two scanning areas, as shown in the figure. Therefore, the search width must be appropriately reduced to exclude the "missed area" and ensure that no area is missed in the search. If point P1 represents the innermost point of the "missed area", and a line parallel to the flight direction is drawn through point P1, then this parallel line is the maximum range that can guarantee no omissions in the search.

[0128] Let d1 represent the coverage width for a search without omissions. The size of d1 is related to factors such as the UAV's flight speed V, the scanning angular velocity ω of the optoelectronic device, the static field of view angle ε, the dynamic search azimuth angle θ1, and the inner / outer radii R1 / R2 of the static search field of view. Multiple factors need to be considered comprehensively to maximize d1.

[0129] First, establish a horizontal two-dimensional coordinate system on the ground, with O1 as the origin, the flight direction as the Y-axis, and the X-axis perpendicular to the Y-axis, with rightward as positive. The trajectory B1B2 and the straight line... The intersection of the two points is the location of point P1. In the coordinate system O1XY, it is represented by (x... P1 ,y P1 Let P1 be a number.

[0130]

[0131] To find point P1, we need to establish the trajectory B1B2 and the line. The mathematical analytical equation.

[0132] At time 0, the aircraft is located at point O1. Let B1 be the starting point of the trajectory. Also within the O1XY horizontal coordinate system, since the field of view undergoes uniform circular motion in the horizontal plane on one hand, and uniform linear flight with the aircraft on the other, the functional expression for trajectory B1B2 is:

[0133]

[0134] x B1 Transform the expression:

[0135] t=(arcsin(x B1 / R2)+θ1) / ω

[0136] x B1 The value range of is [-R2sinθ1, R2sin(θ1-ε)]. Substituting the above equation into y... B1 The expression is:

[0137]

[0138] Similarly, the same method is used to model the trajectory C2C3, with C3 as the starting point and C2 as the ending point. That is, when the trajectory is at point C3, t=0. Then the functional expression of the trajectory C2C3 is:

[0139]

[0140] Similarly, transforming the above equation to eliminate the time variable t, we get:

[0141]

[0142] x C2 The range of values ​​for x is [-R1sinθ1, R1sin(θ1-ε)]. Since R2 > R1, therefore x B1 The range of values ​​for x includes C2 The range of values ​​for .

[0143] straight line The function expression is:

[0144]

[0145] The range of values ​​is [-R2sinθ1,0].

[0146] The equation of the composite trajectory B1B2 and The intersection of the equations is the location of point P1, denoted by (x1, y1). Then we have the system of equations:

[0147]

[0148] (x1, y1) can be solved from this system of equations, and d1 = 2|x1|.

[0149] 2. Constraint Modeling

[0150] The search coverage width of UAVs is limited by many factors, but in practical applications, it is mainly determined by experience or simple calculations. This is not only highly subjective but also makes it difficult to maximize search efficiency, and it does not take into account limitations such as area omissions and operator visual fatigue. This invention focuses on the limitations imposed by speed-to-height ratio and missed scans on coverage width optimization, and establishes corresponding models for each. It also provides other limiting conditions in the form of thresholds or value ranges.

[0151] 2.1 Speed-to-Height Ratio Limitation Model

[0152] To ensure stable imaging of the scene by photoelectric detection equipment, most photoelectric devices have a speed-to-height ratio limitation, such as... Figure 4 Assuming the UAV's flight speed is V and its altitude is h, and the angle of depression formed by any point within the detection area of ​​P2 relative to the location of the photoelectric device is μ′, then μ1 ≤ μ′ ≤ μ2. Let the distance between point P2 and the location of the photoelectric device O1 be R′, then R′ = h / cosμ′, and the angle of deviation from the flight heading be θ′. Assuming the component of velocity V perpendicular to O′P1 is equal to V′, using velocity decomposition and trigonometric functions, V′ can be obtained as follows:

[0153]

[0154] Let ω1 represent the angular velocity of point P1 relative to O′.

[0155] ω1=V′cosμ′ / h

[0156]

[0157] According to the speed-to-height ratio principle, assuming the maximum speed-to-height ratio is γ, then:

[0158]

[0159] Since μ1≤μ′≤μ2, therefore

[0160]

[0161] Therefore, the expression can be replaced with:

[0162]

[0163] Further transforming the left side of the above equation, we get:

[0164]

[0165] Since θ′≤θ1, therefore:

[0166]

[0167] The above equation creates a constraint between "height and velocity V".

[0168] 2.2 No omission search restrictions

[0169] When conditions such as excessive flight speed or insufficient static field of view occur, missed areas may appear between two scanned areas, such as... Figure 5 As shown.

[0170] To avoid this situation, we should ensure that trajectory C2C3 always follows trajectory B1B2. That is, for any x within the interval [-R1sinθ1, R1sin(θ1-ε)] where both trajectory function expressions are applicable, we have:

[0171] y c2 (x)≤y B1 (x)

[0172] This formula will be an important constraint in coverage width planning and search track planning.

[0173] 2.3 Other restrictions

[0174] Due to equipment performance limitations, the field of view depression angle is generally not lower than a certain threshold; otherwise, the equipment will have difficulty tracking the target stably. Let this threshold be μ. min Then we have:

[0175] μ min ≤μ1<μ2

[0176] For flight safety reasons, flight altitudes below a certain value are generally not permitted. A minimum safe flight altitude (h) is typically determined based on aircraft performance and various weather conditions. s The specification, given in a task, is a threshold:

[0177] h≥h s

[0178] Due to the recognition distance requirement, the distance O1 from the outer edge of the static field of view to the drone's position should not exceed [a certain value]. That is This imposes restrictions on the flight altitude of unmanned aircraft. (From equation μ) min ≤μ1<μ2 and μ2=μ1+ε we can know:

[0179] μ2≥μ min +ε

[0180] And flight altitude

[0181]

[0182] Since most current UAVs have long cruising time and range, this application does not consider limitations on range and time. However, the cruising speed of such UAVs is relatively low; assuming the cruising speed can be selected within the range [V1, V2], that is:

[0183] V2≥V≥V1

[0184] When the centerline of the field of view is projected onto the ground and coincides with the direction of the flight velocity V, θ1 = ε / 2; when θ1 = π / 2, the containment width reaches its maximum.

[0185] Due to the requirements of visual recognition and visual fatigue control for task force personnel, the value of ω should not be too large. max The maximum value of ω is:

[0186] ω≤ω max

[0187] 3. Search Parameter Planning Based on Immune Algorithm

[0188] 3.1 Planning Target Model

[0189] For any global optimization problem, it is necessary to establish an optimization index model. However, in many optimization problems, a single parameter cannot be simply used as the optimization index. For example, in the problem in this paper, simply maximizing the coverage width without omissions is insufficient. Instead, the task objective must be used as the final planning objective, and different tasks correspond to different index function models.

[0190] For coverage search, the larger the area covered per unit time, the better. Based on this, an evaluation metric is defined as coverage efficiency η.

[0191] eta=S / t=(V·t·d1) / t=V·d1

[0192] Wherein, is the area searched within time S;

[0193] If the goal is to maximize the coverage area within a given period, then the objective function is:

[0194] max(d1·V)

[0195] The smaller ω is, the better, as it lowers the requirements for the task control personnel and helps them maintain high efficiency and accuracy in the search over a long period of time; therefore, the objective function is: min(ω).

[0196] h0 and V0 represent the optimal cruising speed and altitude under the current weather conditions, respectively, which are more conducive to the aircraft maintaining a longer range. Generally, the optimal speed varies at different altitudes, and the corresponding relationships are complex, which will not be elaborated upon in this paper. Therefore, h0 and V0 are constants in this paper. |Δh0| and |ΔV0| represent the deviations between the flight altitude h and flight speed V of the mission plan and their optimal values, respectively. Therefore, the objective functions are: min(|ΔV0|) and min(|Δh0|).

[0197] Therefore, the objective function for optimizing the task parameters is:

[0198]

[0199] This is a typical multi-objective optimization problem. A common approach to multi-objective optimization problems is the weighted coefficient method. First, the values ​​of each optimization objective are normalized so that they can be compared on the same scale. Then, different weighted coefficients are assigned according to the importance of different optimization objectives. Finally, the weighted sum of the optimization objective functions yields a comprehensive optimization objective function.

[0200] First, the values ​​of each objective function are normalized, resulting in:

[0201]

[0202] Based on the nature of the coverage search task and the relatively long range of typical UAVs, the primary objective is to achieve complete coverage of the target area in the shortest possible time. The next priority is to ensure the search efficiency and accuracy of the task force. Finally, while maintaining the first two objectives, the goal is to maximize economic efficiency. Therefore, the importance of the four objective functions to be optimized is as follows:

[0203] max(d1·V)>min(ω)>min(|ΔV0|)=min(|Δh0|)

[0204] If the weight coefficients of the four objective functions to be optimized are K η K ω , Then it should be guaranteed In this article, K is defined as η =100, K ω =10, The comprehensive optimization objective function J is as follows:

[0205]

[0206] max(J)

[0207] Therefore, the optimized model can be obtained as follows:

[0208]

[0209] max(J)

[0210] T = 4(θ1 - ε / 2) / ω

[0211]

[0212] R1 = h·tan(arctan(R2 / h)-ε)

[0213] μ2 = arctan(R2 / h)

[0214] μ1=μ2-ε

[0215]

[0216] d1 = 2|x1|

[0217]

[0218]

[0219] ST

[0220] θ1≤90°

[0221] V2≥V≥V1

[0222] h≥h s

[0223]

[0224] μ min ≤μ1

[0225] ω≤ω max

[0226]

[0227] y c2 (x)≤y B1 (x), x∈[-R1sinθ1,R1sin(θ1-ε)]

[0228] 3.2 Algorithm Flow

[0229] Global multi-objective optimization is a classic problem. When there are numerous models and constraints, and nonlinear mutations occur, traditional analytical methods struggle to handle them. Evolutionary computation, however, can effectively solve this type of problem and has been widely applied in many applications. This invention uses an immune algorithm with good global optimization capabilities for optimization; the algorithm flow is as follows: Figure 6 As shown.

[0230] Velocity V, flight altitude h, azimuth angle θ1, and scanning angular velocity ω are combined as a set of parameters to be optimized; each set of parameters is considered an antibody in the algorithm population, with a population size of 50. The objective function value J generated based on the set of task parameters is used as the fitness f of the antibody. Clonal selection probability P s =0.4, mutation probability P m =0.8, update probability P u =0.2.

[0231] Step 1: Initialize algorithm-related parameters, initialize search area coordinates, initialize UAV platform performance parameters, initialize optoelectronic device performance parameters, and initialize...

[0232] Step 2: Based on the threshold range of each parameter, randomly generate 50 sets of task parameter groups to form the initial antibody population. Each set of parameters contains 4 parameters: [V,h,θ1,ω].

[0233] Step 3: Calculate the objective function value J for each antibody (a set of task parameters) in the population as the fitness f of that antibody;

[0234] Step 4: Calculate whether each antibody (a set of task parameters) in the population satisfies the speed-to-height ratio constraint and the missed scan constraint. If any one of the constraints is not satisfied, the fitness of the antibody is set to 0 (f=0). If both constraints are satisfied, the fitness value of the antibody remains unchanged.

[0235] Step 5: Determine if the optimal antibody meets the task requirements. If it does, the algorithm ends; otherwise, proceed to the next step.

[0236] Step 6: Calculate the vector moment concentration of each antibody in the population based on fitness, and calculate the selection probability of that antibody based on the vector moment concentration;

[0237] Step 7: Select antibodies and perform clonal amplification based on the concentration regulation mechanism of the immune algorithm, with a clonal selection probability P. s =0.4;

[0238] Step 8: Perform a mutation operation on each antibody in the cloned and amplified population, with a mutation probability P. m =0.8. It should be noted that every parameter of the antibody obtained after mutation must meet the threshold range limit of that parameter. If it does not meet the limit, it must be mutated again.

[0239] Step 9: Determine if the number of antibodies in the current population has reached 50. If not, randomly generate antibodies to replenish the population to 50.

[0240] Step 10: Determine whether the set conditions for the algorithm to run have been met. If they have, the algorithm ends; otherwise, use the new population obtained as the initial population and go to step 3 to start a new round of calculation.

[0241] 3.3 Simulation Experiment

[0242] Simulation calculations will be performed on static field of view, dynamic search range, coverage width without omission, and immune optimization under typical parameter settings to comprehensively evaluate the effectiveness of the established model and optimization method.

[0243] First, set the threshold and range of the parameters. Static field of view ε = 20°, maximum recognition distance d. ShB =10km, minimum depression angle μ min =15°, maximum scanning angular velocity ω max = 8° / s, minimum flight altitude h s =200m, maximum flight altitude h max =8191.5m, minimum speed V1=100km / h, maximum speed V2=300km / h, speed-to-height ratio threshold γ=0.08rad / s.

[0244] Example 1: Static Coverage Area and Search Process Simulation

[0245] First, based on experience, six sets of task parameters were set, and simulations were performed based on the coverage area model established earlier. The specific parameters and their corresponding results are as follows: Figure 7 As shown; the static coverage area based on the first set of parameters is as follows Figure 7 As shown in A1B1C1D1 in (a), the area of ​​the object moves back and forth at a constant speed around its location, with the flight direction as the axis of symmetry. If the mission parameters do not change, its area remains constant during the search process. Trajectory B1B2 is the trajectory of point B1 within half a cycle, and the other trajectories are similar.

[0246] Simulation results show that the model presented in this paper basically reflects the changing characteristics of the static coverage field of view, which varies with height h, pitch angle μ, and maximum recognition distance. The coverage area varies with parameters such as speed-to-height ratio, target recognition, and personnel fatigue. However, a larger static coverage area is not always better; factors such as speed-to-height ratio, target recognition, and personnel fatigue must also be considered. The coverage width and efficiency parameters corresponding to the six sets of task parameters are shown in Table 1.

[0247] Table 1 shows the coverage width and efficiency parameters for the six sets of tasks.

[0248]

[0249] The third set of parameters yields a maximum coverage width of 14979.7m, as shown in the simulation results. Figure 7As shown in (c), the search efficiency is not high at this point; the second set of data shows that decreasing θ can reduce the scan cycle, as shown in (b), while the fourth set of parameters, although the coverage width is not the maximum, can achieve the maximum coverage efficiency of 1175405m. 2 / s, such as Figure 7 As shown in (d), the data from the fifth and sixth groups indicate that lowering the flight altitude can achieve higher coverage efficiency and coverage width, as shown in (e). From (a) and (f), it can be seen that the parameters in the first and sixth groups do not meet the requirement of no-miss coverage, and there are missed areas between their two scans. Therefore, simply relying on empirical settings is not necessarily accurate; as shown in (a) and (f), missed areas exist, and it is difficult to obtain the optimal search width and search efficiency. Therefore, it is necessary to rely on modern optimization theory for parameter optimization.

[0250] Example 2: Coverage Width Immune Optimization Simulation

[0251] In practical tasks, not only is high search efficiency required, but also many other factors need to be considered. Therefore, according to the formula...

[0252]

[0253] The established comprehensive optimization index J is used for optimization research, and K is set. η =100, K ω =10,

[0254] Next, we will observe and compare the changes in the comprehensive optimization index when other parameters are kept constant for different groups.

[0255] like Figure 8 The figure shows the variation of J with V and h when the search azimuth angle θ1 = 85° and the scanning search angular velocity ω = 3° / s. It can be seen from the figure that this correspondence is a curved surface, and when V is large and h is high, there is a large restricted area on the plane formed by V and h, where the parameters do not meet the condition of no omission coverage; when V is large and h is low, there is also a smaller restricted area on the plane formed by V and h.

[0256] like Figure 9 The figure shows the variation of J with azimuth angle θ1 and scanning search angular velocity ω when V = 80 m / s and h = 3500 m. It can be seen from the figure that this correspondence forms a surface. When ω is small and θ1 is large, there is a large restricted region where the parameters do not meet the condition of no omission coverage.

[0257] When V and ω are constant, the changes of J with h and θ1, and the changes of J with V and θ1 when h and ω are constant, are respectively as follows: Figure 10 , Figure 11 As shown, it can be seen that in both cases there is a restricted region and a nonlinear abrupt change.

[0258] Based on the simulation results above, there is no specific proportional relationship between the comprehensive optimization index and a certain task parameter. Furthermore, due to numerous constraints, there is a nonlinear mutation region. Therefore, it is quite difficult to calculate such problems using traditional analytical methods.

[0259] Therefore, coverage width optimization was performed based on an immune algorithm. Using J as the optimization index, the maximum number of generations, Era = 5000, was set, and optimization ended after 5000 generations. Ten rounds of simulation were conducted, and the optimization results are shown in Table 2. It can be seen that the result obtained based on the immune algorithm has a large coverage width and high coverage efficiency, and meets the requirements for no-miss coverage and speed-to-height ratio.

[0260] Table 2. Optimization results after 10 rounds based on the immune algorithm.

[0261]

[0262] However, it can also be seen that there are significant differences between the optimization results of different rounds. Taking the comprehensive index J as an example, the minimum value is 135.01 and the maximum value is 174.54 in the optimization results of 10 rounds. This situation illustrates the complexity of this optimization problem, which is prone to getting trapped in local optima, and also places higher demands on the global optimization algorithm.

[0263] To make the images easier to observe and less crowded, the following illustrations will only show the evolution curves for typical rounds of optimization, rather than the evolution curves for all rounds. Based on the differences in the comprehensive index J of the final optimization results, the optimization results of rounds 2, 5, 7, and 9 are selected as representatives for comparative analysis.

[0264] The curve showing the change of the comprehensive index J in the representative optimization process is as follows: Figure 12 As shown, severe evolutionary stagnation exists in all four optimization processes, indicating that the optimization process gets stuck in local optima and fails to escape. This ultimately leads to significant differences in the comprehensive index J of the optimization results.

[0265] Since d1 is a major component of J, and its weight is much greater than other indicators, the trend of d1 is as follows: Figure 13 The result is largely consistent with J. The difference is that the change of d1 exhibits oscillations, which reflects the constraints on d1. The optimization algorithm needs to consider other parameters in a balanced manner to ensure that the parameter set meets various constraints.

[0266] The rate change curve of a typical optimization process is as follows: Figure 14 As shown in the figure and table, the final speeds of all 10 optimization results are close to the maximum permissible speed of 83.33 m / s (300 km / h), indicating that a higher cruising speed has a positive effect on improving J. Basically, in the early stages of the optimization process, the speed changes exhibit oscillations before stabilizing, as seen in rounds 5, 7, and 9. In the second round of optimization, due to the algorithm getting trapped in a local optimum, the V curve consistently showed abrupt changes, resulting in a poor final result.

[0267] The curves showing the change in height h in the four typical optimization rounds are as follows: Figure 15 As shown in Table 2, the height gradually decreases in all four typical optimization processes, indicating that appropriately reducing the height has a positive effect on improving J. However, a smaller height is not always better. In all 10 optimization rounds, the height h converges to different values. However, the final convergence results of the height parameter obtained from the 10 optimization rounds are inconsistent. Even if the comprehensive index values ​​J of the two sets of parameters are similar, the height values ​​differ significantly. For example, in the results of rounds 5 and 6, the comprehensive index values ​​J are 165.63 and 165.42 respectively, but the height values ​​are 1915.7 and 1303.64 respectively, showing a considerable difference. This reflects the complexity of the optimization problem and confirms the conclusion that a smaller height is not always better.

[0268] The simulation of the scanning trajectory based on typical optimization results and the scanning efficiency curves under the final parameters are shown below. Figure 16 , Figure 17 As shown, under the constraints of various thresholds, the optimized coverage width is generally large, and no missed areas are generated during the scanning process.

[0269] In summary, this invention has conducted in-depth research on the calculation, constraints, and optimization of coverage width when using optoelectronic equipment for area coverage without omissions by UAVs. A calculation model for the coverage width without omissions under static coverage area and typical scanning methods of the optoelectronic equipment was established. Models for coverage width constraints and speed-to-height ratio constraints were also established. Simulation calculations show that the static coverage area conforms to the actual situation, the calculation method for coverage width without omissions is accurate, and the constraints accurately reflect the influence of various factors on the coverage width. Modeling the coverage width and constraints demonstrates that the coverage width model changes non-linearly and the constraint relationships are complex. Simulations show that an immune evolutionary algorithm can effectively solve this problem, and the coverage process based on optimized parameters conforms to various constraints, has higher search efficiency, and does not miss any areas.

[0270] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0271] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for optimizing parameters for a UAV's mission of achieving complete coverage of a region, characterized in that, Includes the following steps: S1. Establish the coverage width model: S11. Determine the scanning mode of the optoelectronic device; S12. Establish a static field-of-view model for optoelectronic equipment; S13. Establish a no-omission coverage width model: S2. Establish a constraint model; S21. Determine the speed-to-height ratio limiting conditions; S22. Determine the search restrictions for no omissions; S23. Determine the limitations on the field of view depression angle, flight altitude, cruise speed, and the angular velocity of the optical axis of the optoelectronic equipment in the horizontal plane. S3. Search parameter planning based on the immune algorithm: S31. Establish an optimization index model; S32. Optimize the target model based on the immune algorithm: The specific steps of the immune algorithm are as follows: speed Flight altitude Azimuth angular velocity of scanning The combination serves as a set of parameters to be optimized; each set of parameters is considered an antibody in the algorithm population, with a population size of 50; the objective function value J generated based on the set of task parameters is used as the fitness of the antibody. Cloning selection probability Probability of mutation Update probability ; Step 1: Initialize algorithm-related parameters, initialize search area coordinates, initialize UAV platform performance parameters, initialize optoelectronic device performance parameters, and initialize... ; Step 2: Based on the threshold range of each parameter, randomly generate 50 sets of task parameter groups to form the initial antibody population. Each set of parameters contains 4 parameters. ; Step 3: Calculate the objective function value J for each antibody in the population as the fitness of that antibody. ; Step 4: Calculate whether each antibody in the population satisfies the velocity-to-high ratio restriction and the missed scan restriction. If any one of the conditions is not met, the fitness of the antibody is set to 0. If both restrictions are met, the fitness value of the antibody remains unchanged. Step 5: Determine if the optimal antibody meets the task requirements. If it does, the algorithm ends; otherwise, proceed to the next step. Step 6: Calculate the vector moment concentration of each antibody in the population based on fitness, and calculate the selection probability of that antibody based on the vector moment concentration; Step 7: Select antibodies and perform clonal amplification based on the concentration regulation mechanism of the immune algorithm, and determine the clonal selection probability. ; Step 8: Perform mutation operations on each antibody in the cloned and amplified population, with the mutation probability... Each parameter of the antibody obtained after mutation must meet the threshold range limit of that parameter; if it does not meet the limit, it must be mutated again. Step 9: Determine if the number of antibodies in the current population has reached 50. If not, randomly generate antibodies to replenish the population to 50. Step 10: Determine whether the set conditions for the algorithm to run have been met. If they have, the algorithm ends; otherwise, use the new population obtained as the initial population and go to step 3 to start a new round of calculation.

2. The method for optimizing parameters for UAV-based no-omission area coverage missions according to claim 1, characterized in that, The specific steps for establishing the static field-of-view model of the optoelectronic device in S12 are as follows: Under current weather conditions, the maximum recognition distance for the target to be searched is [location to be specified]. Its projection position on the ground is The flight altitude is h. The direction of the projection of the central axis of the field of view onto the ground. For the vertical upward direction, the right-hand rule is used to determine it. axis; The size of the field of view. The angle of depression is the inner edge of the field of view. The area formed by ABCD is the coverage area of ​​the static field of view of the optoelectronic device on the ground. The lengths of the field of view sides are respectively... , , As shown in the following formula, ； ； 。 3. The method for optimizing parameters for a drone's no-omission area coverage mission according to claim 1, characterized in that, The specific steps for establishing the no-omission coverage width model in S13 are as follows: by With the origin as the starting point, the flight direction is... axis, Axis perpendicular to Establish a horizontal two-dimensional coordinate system on the ground, with the rightward axis as positive. ； ； in , These are the inner and outer radii of the static coverage area. For flight speed, The rotational angular velocity of the optical axis of the field of view of the optoelectronic device in the horizontal plane; Optoelectronic devices radially along the Y-axis Location moved to When positioning, the scanning optical axis rotates 2 axially from the leftmost position. To the far right, when the photoelectric device moves from... Location moved to When positioning, the scanning optical axis will rotate 2 inch along the axial direction from the rightmost side. The process of returning to the original position constitutes one cycle, and the time taken for one scan cycle is T; within one scan cycle, two circular arc regions are scanned. For dynamic azimuth search, The innermost point of the "missed area" is used. express; The containment width for a complete search.

4. The method for optimizing parameters for UAV-based no-omission area coverage missions according to claim 1, characterized in that, The specific conditions for limiting the medium speed to high ratio in S21 are as follows: ； The angle of deviation from the flight heading is , This represents the maximum speed-to-height ratio.

5. The method for optimizing parameters for a drone's no-omission area coverage mission according to claim 4, characterized in that, The no-omission search constraint in S22 is: The area formed by A1B1C1D1 is where the optoelectronic equipment is located. The static field of view covers the ground area, and the area formed by A2B2C2D2 is where the photoelectric equipment is located. The static field of view covers the ground area, and the area formed by A3B3C3D3 is where the photoelectric equipment is located. The static field of view covers the area on the ground for the trajectory. and trajectory The function expression in the interval any within All of them have: ； Among them, trajectory The function expression is ， ； trajectory The function expression is ， 。 6. The method for optimizing parameters for UAV-based no-omission area coverage missions according to claim 1, characterized in that, The optimization index model in S31 is specifically as follows: ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ， ； in, For coverage efficiency Weighting coefficients Angular velocity of optoelectronic equipment Weighting coefficients For optimal cruising speed Weighting coefficients Optimal cruising altitude The weighting coefficients, , The flight altitudes for the mission plan are respectively Flight speed Deviation from the optimal value; and > > = , To comprehensively optimize the objective function, The angle of depression to the outer edge. The lowest top-down view, For the highest safe flight altitude, For the minimum safe flight altitude, This represents the maximum angular velocity of the optical axis of the optoelectronic device's field of view in the horizontal plane.

Images