A method for extracting bird micro-motion parameters based on laser micro-Doppler echo

Abstract

A method for extracting micro-motion parameters of flying birds based on laser micro-Doppler echoes belongs to the field of laser radar application detection and identification. The method includes the following steps: Step 1: Establishing flapping motion models of flying birds, one-stage flapping-wing drones, and two-stage flapping-wing drones with multiple scattering points; Step 2: Simulating and obtaining the time-frequency spectra of flying birds, one-stage flapping-wing drones, and two-stage flapping-wing drones at different azimuth and pitch angles; Step 3: Extracting micro-motion parameters for identifying the bird motion model and the flapping-wing drone motion model. The invention is a method for extracting micro-motion parameters based on the frequency shift characteristics of the bird's forearm. This method can effectively estimate the length of the bird's forearm, the flapping angle, and the sweep angle, opening up a new technology for the application of laser micro-Doppler detection technology in precise target detection and identification.

Description

Technical Field

[0001] The present invention relates to a method for extracting bird micro-motion parameters based on laser micro-Doppler echoes, and in particular to a method for detecting bird flapping motion based on a coherent laser radar micro-Doppler detection system, belonging to the field of laser radar application detection and identification. Background Art

[0002] With the advancement of drone detection and identification, the biomimetic capabilities of drones have been significantly enhanced, inspired by the exceptional flight techniques of insects and birds in nature, to achieve greater concealment. Flapping-wing systems offer advantages such as high stealth, maneuverability, adaptability, and low noise. However, this also poses significant challenges to public safety and military defense. Most current bird-detecting radar products utilize a combination of S-band horizontal scanning radar and X-band vertical scanning radar. However, these systems have a low detection rate of 75% and exhibit a certain degree of false alarms, making bird detection ineffective.

[0003] Flapping-wing drones, such as the "SmartBird" from Germany's Festo and the "Carrier Pigeon" developed by Northwestern Polytechnical University, can blend in with flocks of birds to deceive microwave radar systems. By integrating deceptive feathers or coatings onto flapping-wing drones, it's even possible to make them indistinguishable from real ones, effectively rendering them "bird spies" undetectable by optical systems and radar. Therefore, distinguishing between birds and flapping-wing drones presents significant challenges, primarily due to the high similarity in their flight patterns and motion characteristics. Flapping-wing drones mimic the flapping of birds' wings, generating lift and propulsion. This makes their micromotion characteristics very similar to those of birds in terms of frequency and amplitude. Furthermore, the signals of flapping-wing drones and birds vary greatly under different flight modes, further complicating differentiation.

[0004] Therefore, how to quickly and accurately identify and distinguish birds from flapping-wing drones based on their motion patterns is a challenging issue in the detection and identification field. To address this, this paper proposes a method for extracting bird micro-motion parameters based on laser micro-Doppler echoes. Due to the higher resolution and detection sensitivity of lidar, this paper utilizes laser micro-Doppler detection technology for the detection and identification of birds and flapping-wing drones. This method extracts characteristic parameters of birds compared to flapping-wing drones and provides a reliable basis for high-precision motion estimation. This method provides a new technical approach for the precise detection and identification of targets using laser micro-Doppler detection and identification technology. Summary of the Invention

[0005] The present invention aims to address the aforementioned issues with the prior art by proposing a method for extracting bird micromotion parameters based on laser micro-Doppler echoes. This method utilizes the time-frequency spectrum of flapping wing micro-Doppler echoes to extract the micromotion parameters needed for identification. This method provides guidance for the detection and identification of birds and flapping-wing drones.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for extracting bird micro-motion parameters based on laser micro-Doppler echoes, the method comprising the following steps:

[0008] Step 1: Establish flapping motion models for birds with multiple scattering points, one-stage flapping-wing drones, and two-stage flapping-wing drones;

[0009] Step 2: Simulate the time-frequency spectra of a flying bird, a one-stage flapping-wing UAV, and a two-stage flapping-wing UAV at different azimuth and pitch angles;

[0010] Step 3: Extract the micro-motion parameters of the flying bird motion model and the flapping-wing drone motion model.

[0011] Furthermore, in step 1, the flapping motion models of the flying bird, the one-stage flapping wing drone, and the two-stage flapping wing drone with multiple scattering points are established; specifically:

[0012] The process of establishing the flapping motion model of a flying bird is as follows: simplifying the bird's wing into two parts: the upper arm and the forearm. The upper arm is connected to the elbow joint through the shoulder joint, and the forearm is connected to the wrist joint through the elbow joint. The elbow joint has only one degree of freedom, and the wrist joint has two degrees of freedom. Based on this, the flapping angle and the sweeping angle of the wing are expressed using sine and cosine functions.

[0013] The process of establishing the flapping motion model of the one-stage flapping-wing drone is as follows: the flapping motion model of the bird is simplified, retaining only the connection between the shoulder joint and the elbow joint, that is, the upper arm, and the elbow joint has only one degree of freedom; and on this basis, the sine and cosine functions are used to represent the flapping angle of the wing;

[0014] The process of establishing the flapping motion model of the two-stage flapping-wing UAV is as follows: the motion pattern of the flapping-wing motion model of the flying bird is simplified, only the connection part between the shoulder joint and the elbow joint, that is, the upper arm; and the connection part between the elbow joint and the wrist joint, that is, the forearm, are retained, and the sweeping motion is ignored, so that the elbow joint and the wrist joint each have only one degree of freedom; and on this basis, the sine and cosine functions are used to represent the flapping angle of the wing.

[0015] Furthermore, in the flapping motion model of the flying bird, the elbow joint has only one degree of freedom, and the wrist joint has two degrees of freedom; specifically, the elbow joint has only one degree of freedom for flapping up and down about the x-axis; the wrist joint has degrees of freedom for flapping up and down and sweeping forward and backward about the mutually perpendicular x-axis and y-axis, respectively;

[0016] In the flapping motion model of the one-stage flapping-wing UAV, the elbow joint has only one degree of freedom; specifically, the elbow joint has only one degree of freedom for flapping up and down about the x-axis;

[0017] In the flapping motion model of the two-stage flapping-wing UAV, the elbow joint and wrist joint each have only one degree of freedom; specifically, the elbow joint has only one degree of freedom for flapping up and down around the x-axis, and the wrist joint has only one degree of freedom for flapping up and down around the x-axis.

[0018] Furthermore, in the flapping motion model of the bird, the sine and cosine functions are used to represent the flapping angle and sweeping angle of the wings, specifically:

[0019] The flapping angle of the wing includes the flapping angle of the upper arm and the flapping angle of the forearm. The flapping angle of the upper arm is expressed by the following formula (1):

[0020] ψ1(t)=A1cos(2πf flap t)+ψ 10 (1)

[0021] Where: A1 represents the maximum flapping angle of the upper arm, f flap represents the flapping frequency, Ψ 10 represents the lag angle of the upper arm flapping, and t represents the time;

[0022] The flapping angle of the forearm is expressed by the following formula (2):

[0023] ψ2(t)=A2cos(2πf flap t)+ψ 20 (2)

[0024] Where: A2 represents the maximum flapping angle of the forearm, f flap represents the flapping frequency, Ψ 20 represents the lag angle of the forearm flapping, and t represents the time;

[0025] The sweep angle of the wing, i.e. the sweep angle of the forearm, is expressed by the following equation (3):

[0026]

[0027] Where: C2 represents the maximum sweep angle of the forearm, f flap represents the flapping frequency, represents the lag angle of the forearm sweep, and t represents the time;

[0028] In the flapping motion model of the one-stage flapping-wing UAV, the sine and cosine functions are used to represent the flapping angle of the wing. The flapping angle of the wing is the flapping angle of the upper arm, which is expressed by formula (1);

[0029] In the flapping motion model of the two-stage flapping-wing UAV, the sine and cosine functions are used to represent the flapping angles of the wings. The flapping angles of the wings include the flapping angles of the upper arm and the forearm. The flapping angle of the upper arm is represented by formula (1); the flapping angle of the forearm is represented by formula (2).

[0030] Furthermore, in step 1, the sweeping frequency of the flying bird motion model is half of the flapping frequency.

[0031] Furthermore, in step 2, the simulation of obtaining the time-frequency spectra of a flying bird, a one-stage flapping-wing drone, and a two-stage flapping-wing drone at different azimuths and pitch angles includes the following steps:

[0032] Step 21: To simulate the time-frequency spectra of the flying bird, the one-stage flapping-wing drone, and the two-stage flapping-wing drone, first derive the common coordinates and final distance formula for the three. The specific derivation process is as follows:

[0033] Define the radar coordinate system as XYZ, and assume that the radar (Q) is located at the coordinate origin O; define the bird coordinate system as xyz, and the center of the upper arm is the coordinate origin o; the bird motion model performs flapping motion around the x-axis. At the initial moment of the bird motion model, it is located at point o (x0, y0, z0). At this time, the distance from the origin o of the bird motion model to the radar is R0, the azimuth angle is α, and the pitch angle is β;

[0034] The direction of the flying bird motion model in the XOY plane is deflected from the positive direction of the X axis to the positive direction of the Y axis, the deflection angle is θ, and the speed is v xy ; The bird motion model moves in the Z-axis direction at a speed of v z The position of the bird's motion model at time t is o′(x0+v xy t·cosθ,y0+v xy t·sinθ,z0+v z t), at this time the distance R from the origin o′ of the bird motion model to the radar (Q) is:

[0035]

[0036] At this time, the azimuth angle α′ and the pitch angle β′ are expressed as:

[0037]

[0038] The echoes of all scattering points on the surface of the bird motion model at time t are expressed as:

[0039]

[0040] Among them, r n (t) is the distance from the scattering point n to the radar (Q) at time t, σ n is the echo intensity coefficient of scattering point n, N is the total number of scattering points on the surface of the bird motion model;

[0041] Micro-Doppler frequency shift f m_d From the phase derivative of the echo model, we can get:

[0042]

[0043] At time t, in order to obtain the echo signal at a certain moment, assuming that a certain scattering point is p, the line of sight vector of the radar (Q) is The distance from the bird motion model to the radar (Q) is recorded as R, and the vector and The angle between the radar (Q) and the scattering point p is γ, assuming that the distance between the radar (Q) and the scattering point p at time t is r p (t);

[0044] The distance modulus r from the scattering point p to the lidar Q is calculated by the triangle cosine theorem p (t):

[0045]

[0046] Where: l p is the distance from the scattering point p to the origin o of the bird motion model, cosγ is the vector and The cosine of the angle between

[0047] The distance formula is simplified to:

[0048]

[0049] vector and The projection cosγ is expressed as:

[0050]

[0051] The final distance formula is expressed as:

[0052]

[0053] Step 22: Simulate and obtain the time-frequency spectrum of a flapping-wing drone. Its position coordinates are:

[0054] The position of the elbow joint is expressed as P1 = [x1, y1, z1], where:

[0055] x1(t)=v·t (13)

[0056] y1(t)=L1 cos[ψ1(t)·π / 180] (14)

[0057] z1(t)=y1(t)·tan[ψ1(t)·π / 180] (15)

[0058] Where: v represents the speed in the horizontal direction toward the radar;

[0059] Step 2 and 3: Simulate and obtain the time-frequency spectrum of the two-stage flapping-wing drone, whose position coordinates are:

[0060] The position of the elbow joint is expressed as P1 = [x1, y1, z1], where:

[0061] x1(t)=v·t (13)

[0062] y1(t)=L1cos[ψ1(t)·π / 180] (14)

[0063] z1(t)=y1(t)·tan[ψ1(t)·π / 180] (15)

[0064] Where: v represents the speed in the horizontal direction toward the radar;

[0065] The position of the wrist joint is expressed as P2 = [x2, y2, z2], where:

[0066] x2(t)=v·t (20)

[0067] y2(t)=L1cos[ψ1(t)·π / 180]+L2·cos[ψ1(t)-ψ2(t)] (21)

[0068] z2(t)=z1(t)+[y2(t)-y1(t)]·tan{[ψ1(t)-ψ2(t)]·π / 180} (22)

[0069] Step 24: Simulate and obtain the time-frequency spectrum of the flying bird. Its position coordinates are:

[0070] The position of the elbow joint is expressed as P1 = [x1, y1, z1], where:

[0071] x1(t)=v·t (13)

[0072] y1(t)=L1cos[ψ1(t)·π / 180] (14)

[0073] z1(t)=y1(t)·tan[ψ1(t)·π / 180] (15)

[0074] Where: v represents the speed in the horizontal direction toward the radar;

[0075] The position of the wrist joint is expressed as P2 = [x2, y2, z2], where:

[0076] x2(t)=[y2(t)-y1(t)]·tan(d) (16)

[0077]

[0078] z2(t)=z1(t)+[y2(t)-y1(t)]·tan{[ψ1(t)-ψ2(t)]·π / 180} (18)

[0079] in:

[0080] Furthermore, in step 2, the simulated wavelength is 1064 nm, the simulated azimuth angle α ranges from 30° to 80°, and the simulated elevation angle β ranges from 20° to 80°.

[0081] Furthermore, in step three, the micro-motion parameters of the bird motion model and the flapping-wing UAV motion model are extracted; the specific steps are:

[0082] Step 31: The motion model of the one-stage flapping-wing UAV relative to the flying bird and the two-stage flapping-wing UAV. In the motion model, the one-stage flapping-wing UAV has no forearm structure, so the parameters of the forearm length and the flapping angle of the forearm are extracted;

[0083] Step 32: The two-stage flapping-wing UAV has no sweeping motion in the motion model relative to the flying bird. Therefore, the sweeping angle of the forearm is parameterized.

[0084] Furthermore, in step 31, the parameters of the forearm length and the forearm flapping angle are extracted; specifically:

[0085] The maximum frequency shift produced by the forearm movement of the flying bird motion model is approximately expressed as:

[0086]

[0087] Get the maximum frequency shift and corresponding time From formula (28), we can get:

[0088]

[0089] Where: represents the maximum frequency shift produced by forearm movement;

[0090] Since there is a lag angle of forearm flapping in the flapping wing motion model, the effective length L of the flapping wing motion model can be approximately expressed as:

[0091] L≈(L1+L2)cosΨ 20 (29)

[0092] Where: 20 represents the lag angle of forearm flapping;

[0093] Based on the above formula, the mathematical relationship between the length of the upper arm, the length of the forearm and the flapping angle of the forearm is approximately obtained, and its expression is:

[0094] LA2≈L2′A2′(30).

[0095] Furthermore, in step 32, the parameters of the sweep angle of the forearm are extracted; specifically:

[0096] Step 321: Use the sweep frequency shift expression to realize the sweep angle of the forearm:

[0097]

[0098] Where: λ is the wavelength, C2 is the sweep angle of the forearm, L2 is the length of the forearm, and f flap represents the flapping frequency, α represents the azimuth angle, β represents the pitch angle, and f C represents the sweep frequency shift;

[0099] Under the condition of known azimuth and elevation angles, the product of L2 and C2 is derived, and its expression is:

[0100]

[0101] Step 322: De-noising to extract accurate frequency information.

[0102] The beneficial effects of the present invention compared to the prior art are: the present invention provides a method for extracting micro-motion parameters of flying birds based on laser micro-Doppler echoes, specifically a method for extracting micro-motion parameters based on the frequency shift characteristics of the flying bird's forearm. This method can effectively estimate the length, flapping angle and sweeping angle of the flying bird's forearm, and opens up a new technology for the application of laser micro-Doppler detection technology in precise target detection and identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 It is the front view of the bird flapping motion model;

[0104] Figure 2 It is a top view of the bird flapping motion model;

[0105] Figure 3 It is a geometric relationship diagram between radar and bird flapping motion model.

[0106] Figure 4 This is a time-frequency diagram of a one-stage flapping-wing UAV.

[0107] Figure 5 It is a time-frequency diagram of flying birds.

[0108] Figure 6 It is a time-frequency diagram of a two-stage flapping-wing UAV.

[0109] The names and reference numerals of the components in the above drawings are as follows:

[0110] Shoulder joint 1, elbow joint 2, wrist joint 3, upper arm 4, forearm 5. DETAILED DESCRIPTION

[0111] This embodiment discloses a method for extracting bird micro-motion parameters based on laser micro-Doppler echoes, the method comprising the following steps:

[0112] Step 1: Establish flapping motion models for birds with multiple scattering points, one-stage flapping-wing drones, and two-stage flapping-wing drones;

[0113] Step 2: Simulate the time-frequency spectra of a flying bird, a one-stage flapping-wing UAV, and a two-stage flapping-wing UAV at different azimuth and pitch angles;

[0114] Step 3: Extract the micro-motion parameters of the flying bird motion model and the flapping-wing drone motion model.

[0115] Furthermore, in step 1, the flapping motion models of the flying bird, the one-stage flapping wing drone, and the two-stage flapping wing drone with multiple scattering points are established; specifically:

[0116] The process of establishing the flapping motion model of the bird is as follows: simplifying the bird's wing into two parts: the upper arm 4 and the forearm 5. The upper arm 4 is connected to the elbow joint 2 via the shoulder joint 1, and the forearm 5 is connected to the wrist joint 3 via the elbow joint 2. The elbow joint 2 has only one degree of freedom, and the wrist joint 3 has two degrees of freedom. On this basis, the sine and cosine functions are used to represent the flapping angle and the sweeping angle of the wing (thereby establishing the flapping motion model of the bird's wing);

[0117] The process of establishing the flapping motion model of the one-stage flapping-wing UAV is as follows: the flapping motion model of the bird is simplified, retaining only the connection between the shoulder joint 1 and the elbow joint 2, i.e., the upper arm 4, and the elbow joint 2 has only one degree of freedom; and on this basis, the sine and cosine functions are used to represent the flapping angle of the wing;

[0118] The process of establishing the flapping motion model of the two-stage flapping-wing UAV is as follows: the motion pattern of the flapping-wing motion model of the flying bird is simplified, only the connection part between the shoulder joint 1 and the elbow joint 2, that is, the upper arm 4; and the connection part between the elbow joint 2 and the wrist joint 3, that is, the forearm 5, are retained, and the sweeping motion is ignored, so that the elbow joint 2 and the wrist joint 3 each have only one degree of freedom; and on this basis, the sine and cosine functions are used to represent the flapping angle of the wing.

[0119] Furthermore, in the flapping motion model of the flying bird, the elbow joint 2 has only one degree of freedom, and the wrist joint 3 has two degrees of freedom; specifically, the elbow joint 2 has only one degree of freedom for flapping up and down about the x-axis; the wrist joint 3 has degrees of freedom for flapping up and down and sweeping forward and backward about the mutually perpendicular x-axis and y-axis, respectively;

[0120] In the flapping motion model of the one-stage flapping-wing UAV, the elbow joint 2 has only one degree of freedom; specifically, the elbow joint 2 has only one degree of freedom for flapping up and down around the x-axis;

[0121] In the flapping motion model of the two-stage flapping-wing UAV, the elbow joint 2 and the wrist joint 3 each have only one degree of freedom; specifically, the elbow joint 2 has only one degree of freedom for flapping up and down around the x-axis, and the wrist joint 3 has only one degree of freedom for flapping up and down around the x-axis.

[0122] Furthermore, in the flapping motion model of the bird, the sine and cosine functions are used to represent the flapping angle and sweeping angle of the wings, specifically:

[0123] The flapping angle of the wing includes the flapping angle of the upper arm 4 and the flapping angle of the forearm 5. The flapping angle of the upper arm 4 is expressed by the following formula (1):

[0124] ψ1(t)=A1cos(2πf flap t)+ψ 10 (1)

[0125] Where: A1 represents the maximum flapping angle of the upper arm 4, f flap represents the flapping frequency, Ψ 10 represents the hysteresis angle of the upper arm 4 flapping, and t represents the time;

[0126] The flapping angle of the forearm 5 is expressed by the following formula (2):

[0127] ψ2(t)=A2cos(2πf flap t)+ψ 20 (2)

[0128] Where: A2 represents the maximum flapping angle of the forearm 5, f flap represents the flapping frequency, Ψ 20represents the hysteresis angle of the forearm 5 flapping, and t represents the time;

[0129] By observing the characteristics of birds performing wing-sweeping motions, we can see that during the flapping of a bird's wings, its forearm 5 simultaneously sweeps. When the wings swing downward, the sweep path at the end of the forearm 5 moves from a position farther from the head to a position closer to the head. When the wings swing upward, the opposite motion occurs. Therefore, the sweep frequency can be defined as half the flapping frequency. This is because a complete flapping cycle consists of one downward and one upward swing, and each swing corresponds to a complete sweeping motion cycle. This definition allows us to more accurately quantify the sweeping frequency of birds, providing a precise parameter basis for further research and analysis.

[0130] The sweep angle of the wing, i.e. the sweep angle of the forearm 5, is expressed by the following formula (3):

[0131]

[0132] Where: C2 represents the maximum sweep angle of the forearm 5, f flap represents the flapping frequency, represents the lag angle of the forearm 5 sweep, and t represents the time;

[0133] In the flapping motion model of the one-stage flapping-wing UAV, the sine and cosine functions are used to represent the flapping angle of the wing. The flapping angle of the wing is the flapping angle of the upper arm 4, which is expressed by formula (1);

[0134] In the flapping motion model of the two-stage flapping-wing UAV, the sine and cosine functions are used to represent the flapping angles of the wings. The flapping angles of the wings include the flapping angles of the upper arm 4 and the forearm 5. The flapping angle of the upper arm 4 is represented by formula (1); the flapping angle of the forearm 5 is represented by formula (2).

[0135] Furthermore, in step 1, the sweeping frequency of the flying bird motion model is half of the flapping frequency.

[0136] Furthermore, in step 2, the simulation of obtaining the time-frequency spectra of a flying bird, a one-stage flapping-wing UAV, and a two-stage flapping-wing UAV at different azimuths and pitch angles includes the following steps:

[0137] Step 21: To simulate the time-frequency spectra of the flying bird, the one-stage flapping-wing drone, and the two-stage flapping-wing drone, first derive the common coordinates and final distance formula for the three. The specific derivation process is as follows:

[0138] Define the radar coordinate system as XYZ, and assume that the radar Q is located at the coordinate origin O; define the bird coordinate system as xyz, and the center of the upper arm 4 is the coordinate origin o; for the convenience of analysis, the direction of the ox axis of the bird coordinate system is the same as the direction of the OX axis of the radar coordinate system, the direction of the oz axis is the same as the direction of the OZ axis of the radar coordinate system, and the directions of the oy axis and the OY axis are determined according to the right-hand rule. The geometric relationship between the radar Q and the flapping wing model is as follows: Figure 3 shown.

[0139] The bird motion model flaps around the x-axis. At the initial moment, it is located at point o(x0, y0, z0). At this time, the distance from the origin o of the bird motion model to the radar is R0, the azimuth angle is α, and the pitch angle is β.

[0140] The direction of the flying bird motion model in the XOY plane is deflected from the positive direction of the X axis to the positive direction of the Y axis, the deflection angle is θ, and the speed is v xy ; The bird motion model moves in the Z-axis direction at a speed of v z The position of the bird's motion model at time t is o′(x0+v xy t·cosθ,y0+v xy t·sinθ,z0+v z t), at this time the distance R from the origin o′ of the bird motion model to the radar Q is:

[0141]

[0142] At this time, the azimuth angle α′ and the pitch angle β′ are expressed as:

[0143]

[0144] Due to the short observation time (generally about 1 second), the position of the bird motion model does not change much, and at this time α′≈α, β′≈β;

[0145] The echoes of all scattering points on the surface of the bird motion model at time t are expressed as:

[0146]

[0147] Among them, r n (t) is the distance from the scattering point n to the radar Q at time t, σ n is the echo intensity coefficient of scattering point n, N is the total number of scattering points on the surface of the bird motion model;

[0148] Micro-Doppler frequency shift f m_d From the phase derivative of the echo model, we can get:

[0149]

[0150] At time t, to obtain the echo signal at a certain moment, assume that a certain scattering point is p, and the line-of-sight vector of radar Q is The distance from the bird motion model to radar Q is denoted as R, and the vector and The included angle is γ. Assume that the distance between radar Q and scattering point p at time t is r p (t);

[0151] The distance modulus r from the scattering point p to lidar Q is calculated by the cosine theorem of a triangle p (t):

[0152]

[0153] In the formula: l p is the distance from the scattering point p to the origin o of the bird motion model, and cosγ is the cosine of the included angle between the vector and ;

[0154] Under normal circumstances, the size of the bird motion model is much smaller than the distance between it and the radar. It can be seen that l p <<R0, and the distance formula is simplified to:

[0155]

[0156] The projection cosγ of the vector and is expressed as:

[0157]

[0158] Then the final distance formula is expressed as:

[0159]

[0160] Step two: Simulate the time-frequency spectrum of a one-piece flapping-wing UAV, and its position coordinates are:

[0161] The position of the elbow joint 2 is expressed as P1 = [x1, y1, z1], where:

[0162] x1(t) = v·t (13)

[0163] y1(t) = L1cos[ψ1(t)·π / 180] (14)

[0164] z1(t) = y1(t)·tan[ψ1(t)·π / 180] (15)

[0165] Among them: v represents the speed flying towards the radar in the horizontal direction;

[0166] Step 2 and 3: Simulate and obtain the time-frequency spectrum of the two-stage flapping-wing drone, whose position coordinates are:

[0167] The position of elbow joint 2 is expressed as P1 = [x1, y1, z1], where:

[0168] x1(t)=v·t (13)

[0169] y1(t)=L1cos[ψ1(t)·π / 180] (14)

[0170] z1(t)=y1(t)·tan[ψ1(t)·π / 180] (15)

[0171] Where: v represents the speed in the horizontal direction toward the radar;

[0172] The position of wrist joint 3 is expressed as P2 = [x2, y2, z2], where:

[0173] x2(t)=v·t (20)

[0174] y2(t)=L1cos[ψ1(t)·π / 180]+L2·cos[ψ1(t)-ψ2(t)] (21)

[0175] z2(t)=z1(t)+[y2(t)-y1(t)]·tan{[ψ1(t)-ψ2(t)]·π / 180} (22)

[0176] Step 24: Simulate and obtain the time-frequency spectrum of the flying bird. Its position coordinates are:

[0177] The position of elbow joint 2 is expressed as P1 = [x1, y1, z1], where:

[0178] x1(t)=v·t (13)

[0179] y1(t)=L1cos[ψ1(t)·π / 180] (14)

[0180] z1(t)=y1(t)·tan[ψ1(t)·π / 180] (15)

[0181] Where: v represents the speed in the horizontal direction toward the radar;

[0182] The position of wrist joint 3 is expressed as P2 = [x2, y2, z2], where:

[0183] x2(t)=[y2(t)-y1(t)]·tan(d) (16)

[0184]

[0185] z2(t)=z1(t)+[y2(t)-y1(t)]·tan{[ψ1(t)-ψ2(t)]·π / 180} (18)

[0186] in:

[0187] Furthermore, in step 2, the simulated wavelength is 1064 nm, the simulated azimuth angle α ranges from 30° to 80°, and the simulated elevation angle β ranges from 20° to 80°.

[0188] Furthermore, in step three, the micro-motion parameters of the bird motion model and the flapping-wing drone motion model are extracted; the specific steps are as follows:

[0189] Step 31: The one-stage flapping-wing UAV has no forearm 5 structure in the motion model compared to the two-stage flapping-wing UAV. Therefore, the length of the forearm 5 and the flapping angle of the forearm 5 are parameterized.

[0190] The micro-motion parameter extraction of the one-stage flapping wing drone in the motion model is compared with the two-stage flapping wing drone. Currently, most flapping wing drone systems use a one-stage structure with only the upper arm 4. This design is significantly different from the real bird's wing in structure. By simplifying the flapping wing model, the spectrum of the one-stage flapping wing drone can be obtained, such as Figure 4 and Figure 5 As shown in the figure, it can be clearly seen from the spectrum analysis that the spectrum of this type of flapping-wing UAV has two equal peaks and only one outer contour. Therefore, by extracting the micro-motion parameters of the forearm 5, the structural identification micro-motion parameters can be obtained.

[0191] Step 32: The two-stage flapping wing UAV has no sweeping motion in the motion model relative to the flying bird. Therefore, the sweeping angle of the forearm 5 is parameterized.

[0192] Two-stage flapping-wing drones extract micro-motion parameters from the motion model of flying birds. To better resemble real birds, some flapping-wing drones adopt a two-stage design. Structurally, this design is not significantly different from real birds. However, in terms of movement, the wing motion of these flapping-wing drones is mostly concentrated on simple up-and-down flapping movements, which is significantly different from the flexible and complex flapping movements of real birds.

[0193] pass Figure 6It can be seen that when the flapping motion component has zero radial velocity relative to the radar, the time-frequency plot shows a sweeping frequency shift. Under the current simulation settings, at t = nT, where n = 0, 1, 2, 3…k, where k is an integer, the sweeping frequency shift caused by the sweeping motion can be determined by analyzing the expression for the maximum frequency shift of the left and right elbow joints.

[0194] Keeping other conditions unchanged, as the sweep angle of forearm 5 increases, the bandwidth of the sweep frequency shift gradually expands. This is because the increase in the sweep angle of forearm 5 causes a change in the motion trajectory, which in turn affects the amplitude and range of the sweep frequency shift.

[0195] Furthermore, in step 31, the parameters of the length of the forearm 5 and the flapping angle of the forearm 5 are extracted; specifically:

[0196] The maximum frequency shift produced by the forearm 5 movement of the flying bird motion model is approximately expressed as:

[0197]

[0198] Get the maximum frequency shift and corresponding time From formula (28), we can get:

[0199]

[0200] Where: represents the maximum frequency shift produced by the movement of forearm 5;

[0201] Since there is a lag angle of the forearm 5 during the flapping motion model, the effective length L of the flapping motion model can be approximately expressed as:

[0202] L≈(L1+L2)cosΨ 20 (29)

[0203] Where: 20 represents the hysteresis angle of the flapping of forearm 5;

[0204] Based on the above formula, the mathematical relationship between the length of the upper arm 4, the length of the forearm 5 and the flapping angle of the forearm 5 is approximately obtained, and the expression is:

[0205] LA2≈L2′A2′(30).

[0206] Furthermore, in step 32, the parameters of the sweep angle of the forearm 5 are extracted; specifically:

[0207] Step 321: The sweep angle of forearm 5 is realized using the expression of sweep frequency shift:

[0208]

[0209] Where: λ represents the wavelength, C2 represents the sweep angle of the forearm 5, L2 represents the length of the forearm 5, and f flap represents the flapping frequency, α represents the azimuth angle, β represents the pitch angle, and f C represents the sweep frequency shift;

[0210] Under the condition of known azimuth and elevation angles, the product of L2 and C2 is derived, and its expression is:

[0211]

[0212] Step 322: De-noising (since the acquired time-frequency diagram contains noise, de-noising is required) to extract accurate frequency information.

[0213] Example 1:

[0214] like Figure 1 、 Figure 2 As shown, this embodiment discloses a method for extracting bird micro-motion parameters based on laser micro-Doppler echo, the method comprising the following steps:

[0215] Step 1: Construct a flapping motion model of a flying bird, with the x-axis being the motion axis. The motion parameters of the model are shown in Table 1.

[0216] Table 1

[0217]

[0218] According to the structural characteristics of the current flapping-wing UAV, Figure 1 、 Figure 2 The flying bird motion diagram shown is simplified by retaining only the connection between the shoulder joint 1 and the elbow joint 2 to obtain the motion model of a one-stage flapping-wing drone.

[0219] According to the current motion characteristics of flapping-wing UAV, the motion model of a two-stage flapping-wing UAV can be obtained by simplifying the flying bird motion pattern and ignoring the sweeping wing motion.

[0220] Step 2: Because the structure and motion of a flapping-wing drone are simplified versions of a flying bird, we first study the complete flapping-wing motion model. Through in-depth analysis of this complete flapping-wing motion model, we can fully understand the fundamental laws and characteristics of flapping-wing motion. Based on this foundation, we further simplify the flapping-wing motion model to capture the motion characteristics of a flapping-wing drone. The following study only considers flapping-wing motion. The radar simulation parameters are shown in Table 2.

[0221] Table 2

[0222]

[0223] According to formula (8), when the emission wavelength is constant, the micro-Doppler shift is primarily related to the distance from the target to the radar. The time-frequency characteristics obtained through laser detection and information processing are the time-frequency characteristics of the superposition of the scattering points on the target surface. Because the outer contour peak of the time-frequency characteristics of the laser frequency band represents the micro-Doppler peak at the vertex, the joint connection points, the left elbow joint P1 and the left wrist joint P2, are selected for calculation.

[0224]

[0225] The distance between the left elbow joint point P1 and the radar can be expressed as:

[0226]

[0227] in: R represents the distance from the target to the radar, α represents the azimuth angle, β represents the pitch angle, represents the radar’s sight vector,

[0228] The calculation formula of the micro-Doppler frequency shift of the left elbow joint point P1 can be expressed as:

[0229]

[0230] in: λ represents wavelength, v xy represents the speed of the flapping target in the XOY plane, θ represents the deflection angle, and v z represents the speed of the flapping target in the Z-axis direction, L1 represents the length of the upper arm 4, A1 represents the maximum flapping angle of the upper arm 4, and f flap represents the flapping frequency, and t represents the time;

[0231] The frequency shift caused by translation can be expressed as:

[0232]

[0233] The distance between the left wrist joint point P2 and the radar can be expressed as:

[0234]

[0235] in:

[0236] The calculation formula of the micro-Doppler frequency shift of the left wrist joint point P2 can be expressed as:

[0237]

[0238] Where: ψ1 and ψ2 represent the flapping angles of the upper arm 4 and the forearm 5, respectively. represents the micro-Doppler frequency shift of the left elbow joint, L2 represents the length of the forearm 5, represents the sweep angle of the forearm 5;

[0239] The position of the left wrist joint 5 is expressed as P2 = [x2, y2, z2], where:

[0240] x2(t)=v·t (20)

[0241] y2(t)=L1cos[ψ1(t)·π / 180]+L2·cos[ψ1(t)-ψ2(t)] (21)

[0242] z2(t)=z1(t)+[y2(t)-y1(t)]·tan{[ψ1(t)-ψ2(t)]·π / 180} (22)

[0243] Similarly, the frequency shift of the right elbow joint and right wrist joint can be obtained as:

[0244]

[0245] Step 3: Extract the micro-motion parameters of the flying bird motion model and the flapping-wing drone motion model.

[0246] Step 31: Extract the micro-motion parameters of the one-stage flapping-wing UAV relative to the flying bird motion model and the two-stage flapping-wing UAV on the motion model. Figure 4 and Figure 5 It can be seen that the time-frequency diagram of the one-stage flapping-wing UAV is relatively simple, without the frequency shift caused by the movement of the forearm 5.

[0247] The maximum frequency shift expression generated by the forearm 5 movement of the flying bird motion model can be approximately expressed as:

[0248]

[0249] Obtain the maximum frequency shift and corresponding time through the time-frequency diagram From formula (28), we can get:

[0250]

[0251] Where: represents the maximum frequency shift produced by the movement of forearm 5;

[0252] Since there is a flapping lag angle of the forearm 5 during the flapping motion model, the effective length L of the flapping motion model can be approximately expressed as:

[0253] L≈(L1+L2)cosΨ 20 (29)

[0254] Where: 20represents the hysteresis angle of the flapping of forearm 5;

[0255] Based on the above formula, the mathematical relationship between the length of the upper arm 4, the length of the forearm 5 and the flapping angle of the forearm 5 can be approximately obtained, and the expression is:

[0256] LA2≈L2′A2′(30)

[0257] When the signal-to-noise ratio SNR=5, the estimated relationship between the length of the upper arm 4, the length of the forearm 5 and the flapping angle of the forearm 5 of the two-stage flapping-wing drone is 0.155 (the theoretical value is 0.16).

[0258] Step 32: Extract the micro-motion parameters of the two-stage flapping wing drone relative to the flying bird on the motion model, through Figure 5 and Figure 6 It can be seen that the flapping motion of a flying bird will produce a sweeping frequency shift. By analyzing formulas (24) and (25), the sweeping frequency shift produced by the sweeping motion can be obtained. Its mathematical expression is:

[0259]

[0260] Under the condition of known azimuth and elevation angles, the product of L2 and C2 can be derived, and its expression is:

[0261]

[0262] Since the acquired time-frequency diagram contains noise, the following processing steps are required to extract accurate frequency information:

[0263] First, based on the estimated flapping frequency, the signal is clipped for one period. The resulting time-frequency plot is then subjected to median filtering and brightening. Median filtering effectively removes isolated noise points, while brightening enhances the visualization of valid signals. Furthermore, the processed time-frequency plot is subjected to dimensionality reduction and point-based processing to further reduce the impact of noise and preserve the signal's key features.

[0264] Next, we perform singular value decomposition (SVD) on the processed time-frequency graph to decompose the original data into multiple singular values and corresponding singular vectors. Due to the presence of noise, the decomposed frequency values will be mixed with noise components. By setting a threshold to limit the size of frequency clusters, we can filter out most of the noise frequency clusters.

[0265] Then, based on the effective frequency after filtering out the noise frequency clusters, the frequency range is restricted to remove the noise in the time-frequency graph after dimensionality reduction. Finally, the sweep frequency shift is obtained from the time-frequency graph after dimensionality reduction, thereby obtaining the product of L2 and C2.

[0266] When the signal-to-noise ratio SNR=5, the product of the length of the forearm 5 and the sweep angle of the forearm 5 is 0.0682 (the theoretical value is 0.0698).

[0267] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for extracting bird micro-motion parameters based on laser micro-Doppler echo, characterized by: The method comprises the following steps: Step 1: Establish flapping motion models for birds with multiple scattering points, one-stage flapping-wing drones, and two-stage flapping-wing drones; Step 2: Simulate the time-frequency spectra of a flying bird, a one-stage flapping-wing UAV, and a two-stage flapping-wing UAV at different azimuth and pitch angles; Step 3: Extract the micro-motion parameters of the bird motion model and the flapping-wing UAV motion model. The specific steps are as follows: Step 31: The one-stage flapping wing drone has no forearm (5) structure in terms of the motion model relative to the flying bird and the two-stage flapping wing drone. Therefore, the length of the forearm (5) and the flapping angle of the forearm (5) are parameterized. Specifically: The maximum frequency shift produced by the forearm (5) motion of the flying bird motion model can be approximately expressed as: (27) Where: , represents the wavelength, represents the flapping frequency, Indicates time; , , is the azimuth, is the pitch angle; Get the maximum frequency shift and corresponding time , obtained from formula (28): (28) Where: represents the maximum frequency shift produced by the movement of the forearm (5); Since the flapping motion model has a lag angle of the forearm (5) during the movement, the effective length of the flapping motion model is It can be expressed approximately as: (29) Where: represents the lag angle of the flapping of the forearm (5), represents the length of the upper arm (4), represents the length of the forearm (5); Based on the above formula, the mathematical relationship between the length of the upper arm (4), the length of the forearm (5) and the flapping angle of the forearm (5) is approximately obtained, and the expression is: (30)； Where: represents the maximum flapping angle of the forearm (5); Step 32: The two-stage flapping wing UAV has no sweeping motion in the motion model relative to the flying bird. Therefore, the sweeping angle of the forearm (5) is parameterized. Specifically: Step 321: The sweep angle of the forearm (5) is realized using the expression of sweep frequency shift: （31） Where: represents the wavelength, represents the sweep angle of the forearm (5), represents the length of the forearm (5), represents the flapping frequency, represents the azimuth, represents the pitch angle, represents the sweep frequency shift; Under the condition of known azimuth and elevation angles, we can deduce and The product of is expressed as: (32) Step 322: De-noising to extract accurate frequency information.

2. The method for extracting bird micro-motion parameters based on laser micro-Doppler echo according to claim 1, characterized in that: In step 1, the flapping wing motion models of the flying bird, the one-stage flapping wing UAV and the two-stage flapping wing UAV with multiple scattering points are established; specifically: The process of establishing the flapping motion model of the flying bird is as follows: simplifying the bird's wing into two parts, the upper arm (4) and the forearm (5), wherein the upper arm (4) is connected to the elbow joint (2) via the shoulder joint (1), and the forearm (5) is connected to the wrist joint (3) via the elbow joint (2), wherein the elbow joint (2) has only one degree of freedom and the wrist joint (3) has two degrees of freedom; and on this basis, using sine and cosine functions to represent the flapping angle and the sweeping angle of the wing; The process of establishing the flapping wing motion model of the one-stage flapping wing UAV is as follows: simplifying the flapping wing motion model of the flying bird, retaining only the connecting part between the shoulder joint (1) and the elbow joint (2), that is, the upper arm (4), and the elbow joint (2) has only one degree of freedom; and on this basis, using sine and cosine functions to represent the flapping angle of the wing; The process of establishing the flapping wing motion model of the two-stage flapping wing UAV is as follows: simplifying the motion pattern of the flapping wing motion model of the flying bird, retaining only the connecting part between the shoulder joint (1) and the elbow joint (2), i.e., the upper arm (4); and the connecting part between the elbow joint (2) and the wrist joint (3), i.e., the forearm (5), ignoring the sweeping motion, so that the elbow joint (2) and the wrist joint (3) each have only one degree of freedom; and on this basis, using sine and cosine functions to represent the flapping angle of the wing.

3. The method for extracting bird micro-motion parameters based on laser micro-Doppler echo according to claim 2, characterized in that: In the flapping motion model of the flying bird, the elbow joint (2) has only one degree of freedom, and the wrist joint (3) has two degrees of freedom; specifically, the elbow joint (2) has only one degree of freedom for flapping up and down around the x-axis; the wrist joint (3) has the degrees of freedom for flapping up and down and sweeping forward and backward around the mutually perpendicular x-axis and y-axis, respectively; In the flapping wing motion model of the one-stage flapping wing UAV, the elbow joint (2) has only one degree of freedom; specifically, the elbow joint (2) has only one degree of freedom for flapping up and down around the x-axis; In the flapping wing motion model of the two-stage flapping wing UAV, the elbow joint (2) and the wrist joint (3) each have only one degree of freedom; specifically, the elbow joint (2) has only one degree of freedom for flapping up and down around the x-axis, and the wrist joint (3) has only one degree of freedom for flapping up and down around the x-axis.

4. The method for extracting bird micro-motion parameters based on laser micro-Doppler echo according to claim 2, characterized in that: In the flapping motion model of the bird, the sine and cosine functions are used to represent the flapping angle and sweeping angle of the wings, specifically: The flapping angle of the wing includes the flapping angle of the upper arm (4) and the flapping angle of the forearm (5). The flapping angle of the upper arm (4) is expressed by the following formula (1): （1） in: represents the maximum flapping angle of the upper arm (4), represents the flapping frequency, represents the hysteresis angle of the upper arm (4) flapping, Indicates time; The flapping angle of the forearm (5) is expressed by the following formula (2): （2） in: represents the maximum flapping angle of the forearm (5), represents the flapping frequency, represents the hysteresis angle of the forearm (5) flapping, Indicates time; The sweep angle of the wing, i.e. the sweep angle of the forearm (5), is expressed by the following formula (3): （3） in: represents the maximum sweep angle of the forearm (5), represents the flapping frequency, represents the lag angle of the forearm (5) sweep, Indicates time; In the flapping motion model of the one-stage flapping-wing UAV, the sine and cosine functions are used to represent the flapping angle of the wing. The flapping angle of the wing is the flapping angle of the upper arm (4), which is represented by formula (1); In the flapping motion model of the two-stage flapping-wing UAV, the sine and cosine functions are used to represent the flapping angles of the wings. The flapping angles of the wings include the flapping angles of the upper arm (4) and the flapping angles of the forearm (5). The flapping angles of the upper arm (4) are represented by formula (1); and the flapping angles of the forearm (5) are represented by formula (2).

5. The method for extracting bird micro-motion parameters based on laser micro-Doppler echo according to claim 1, characterized in that: In step 1, the sweeping frequency of the bird motion model is half of the flapping frequency.

6. The method for extracting bird micro-motion parameters based on laser micro-Doppler echo according to claim 1, characterized in that: In step 2, the simulation of obtaining the time-frequency spectra of a flying bird, a one-stage flapping-wing UAV, and a two-stage flapping-wing UAV at different azimuths and pitch angles includes the following steps: Step 21: To simulate the time-frequency spectra of the flying bird, the one-stage flapping-wing drone, and the two-stage flapping-wing drone, first derive the common coordinates and final distance formula for the three. The specific derivation process is as follows: Define the radar coordinate system as XYZ, and assume that the radar Q is located at the coordinate origin O; define the bird coordinate system as xyz, and the center of the upper arm (4) is the coordinate origin o; the bird motion model performs flapping motion around the x-axis, and at the initial moment of the bird motion model, it is located at point At this time, the distance from the origin o of the bird motion model to the radar is , the azimuth is , the pitch angle is ; The direction of motion of the flying bird motion model in the XOY plane deflects from the positive direction of the X axis to the positive direction of the Y axis, and the deflection angle is , the speed is ; The bird motion model moves in the Z-axis direction at a speed If the bird moves, the position of the bird motion model at time t is , at this time the origin of the bird motion model Distance from point to radar Q for: (4) At this time, the azimuth and pitch angle Expressed as: (5) (6) The echoes of all scattering points on the surface of the bird motion model at time t are expressed as: (7) in, is the distance from the scattering point n to the radar Q at time t, is the echo intensity coefficient of scattering point n, N is the total number of scattering points on the surface of the bird motion model; Micro-Doppler shift From the phase derivative of the echo model, we can get: (8) At time t, in order to obtain the echo signal at a certain moment, assuming that a certain scattering point is p, the line of sight vector of radar Q is , the distance from the bird motion model to the radar Q is recorded as ,vector and The angle is , assuming that the distance between radar Q and scattering point p at time t is ; The distance modulus from the scattering point p to the lidar 𝑄 is calculated by the triangle cosine theorem : (9) Where: is the distance from the scattering point p to the origin o of the bird motion model, is a vector and The cosine of the angle between The distance formula is simplified to: (10) vector and Projection Expressed as: (11) The final distance formula is expressed as: (12)； Step 22: Simulate and obtain the time-frequency spectrum of a flapping-wing drone. Its position coordinates are: The position of the elbow joint (2) is expressed as ,in: (13) (14) (15) Where: v represents the speed in the horizontal direction toward the radar; Step 2 and 3: Simulate and obtain the time-frequency spectrum of the two-stage flapping-wing drone, whose position coordinates are: The position of the elbow joint (2) is expressed as ,in: (13) (14) (15) Where: v represents the speed in the horizontal direction toward the radar; The position of the wrist joint (3) is expressed as ,in: （20） （21） （22） Step 24: Simulate and obtain the time-frequency spectrum of the flying bird. Its position coordinates are: The position of the elbow joint (2) is expressed as ,in: (13) (14) (15) Where: v represents the speed in the horizontal direction toward the radar; The position of the wrist joint (3) is expressed as ,in: (16) (17) (18) in: .

7. The method for extracting bird micro-motion parameters based on laser micro-Doppler echo according to claim 6, characterized in that: In step 2, the simulated wavelength is 1064nm and the simulated azimuth is The range is 30°~80°, the pitch angle The range is 20°~80°.

Images