Unmanned aerial vehicle blade geometric parameter estimation method based on time-frequency peak value time shifting

Abstract

The invention discloses an unmanned aerial vehicle blade geometric parameter estimation method based on time-frequency peak value time shifting, and belongs to the field of laser radar application detection. The method comprises the following steps: step 1, constructing a multi-point scattering single-blade echo model of the rotor unmanned aerial vehicle; step 2, simulating a single-blade echo signal of the rotor; 3, rotor blade width estimation and error analysis based on frequency peak value time shifting; and 4, rotor blade length-width ratio estimation and error analysis based on frequency peak value time shifting. According to the method, high-recognition-rate extraction of characteristic parameters such as the width and the length-width ratio of the rotor wing is achieved through peak value time shifting characteristics by means of a large amount of rotor wing detail information contained in high-spatial-resolution laser radar echo signals. According to the method, the width and the length-width ratio of the unmanned rotorcraft can be effectively estimated, a theoretical basis is provided for high-accuracy identification of the unmanned rotorcraft, and technical support is provided for high-precision identification of the unmanned rotorcraft.

Description

Technical Field

[0001] The present invention belongs to the field of laser radar application detection, and specifically relates to a method for estimating geometric parameters of unmanned aerial vehicle blades based on time-frequency peak time shift. Background Art

[0002] The rotor is the main power component of the rotary-wing UAV. Its rotational motion modulates the Doppler echo generated by the motion of the UAV fuselage to form a micro-Doppler signal. The echo signal contains the size information of the rotor. Different types of rotary-wing UAVs have different blade parameter information, mainly including blade width and the aspect ratio of the rotor blades, which are important reference indicators for identifying rotary-wing UAVs. At present, the parameters for UAV identification mainly include rotor rotation frequency, rotor length, number of blades, and number of rotors. The rotor rotation frequency is mainly identified by the number of cycles completed by the rotor per unit time. Since a single parameter cannot completely and accurately identify the model of the UAV, in order to further improve the accuracy of identifying the model of the UAV, different frequencies of rotor rotation during flight have different echo spectra. Although the number of rotor rotations can be extracted based on the echo spectrum characteristics, the rotor length cannot be extracted based on this method; the number of blades is derived from the number of flashes in the time-frequency diagram, but this parameter is limited to microwave radar and only a few rotor tip points can be taken when modeling. At the same time, a large amount of rotor detail information is lost due to the small number of points taken. Therefore, there is an urgent need for a high-precision recognition method that can accurately identify the geometric parameters (blade width and aspect ratio) of UAVs. Summary of the invention

[0003] The purpose of the present invention is to solve the above-mentioned problems existing in the background technology and to propose a method for estimating geometric parameters of UAV blades based on time-frequency peak time shift.

[0004] The present invention utilizes a large amount of rotor detail information contained in the high spatial resolution lidar echo signal and realizes high recognition rate extraction of characteristic parameters such as rotor width and aspect ratio through peak time shift features.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A method for estimating geometric parameters of UAV blades based on time-frequency peak time shift, the method comprising the following steps:

[0007] Step 1: Construction of the multi-point scattering single blade echo model of the rotorcraft UAV;

[0008] Step 2: Simulation of the echo signal of a single rotor blade;

[0009] Step 3: Estimation and error analysis of rotor blade width based on peak frequency time shift;

[0010] Step 4: Estimation and error analysis of rotor blade aspect ratio based on frequency peak time shift.

[0011] Furthermore, in step 1, the construction of the multi-point scattering single blade echo model of the rotor UAV is specifically as follows:

[0012] The rotor blade rotation center is established and located at the coordinate origin O, the azimuth angle α = 0°, the pitch angle is β, and the distance from the laser radar to the rotor center is R 0 , the rotor blade height is z 0 , the distance from the scattering point on the rotor blade to the rotor center is l M ,R 0 >>l M , the scattering point rotates around the Z axis at a constant rotation speed ω; let the initial phase of point M on the rotor blade be The rotor blade reaches point M' after time t, and its phase is

[0013] From the single-point modeling, we know that the baseband signal returned by the scattering point M' at a distance R is:

[0014]

[0015] Where: f 0 represents the carrier frequency of the laser radar, λ represents the wavelength, and R M (t) represents the distance between point M on the constructed rotor blade model and the radar at time t, and σ represents the scattering rate;

[0016] The phase term is derived to obtain the micro-Doppler frequency shift caused by the rotation of point M on the rotor blade, that is:

[0017]

[0018] in: represents phase, t represents time, λ represents wavelength, and Δt represents peak time shift;

[0019] The rotor blade is equivalent to a cuboid model with thickness, and the cuboid model is meshed. The points on the mesh are called scattering points, which replace the center point of the rotor blade scattering surface element. Let the coordinates of each scattering point be (x i ,y j ,z k ), where i, j, k are integers; the distance from the rotor center O to each scattering point is l M (x,y,z)=(x 2 +y 2 +z 2 ) 1 / 2 , the rotor blade echo signal is:

[0020]

[0021] Where: σ represents the scattering rate, f 0 represents the carrier frequency of the laser radar, R M (t) represents the distance between point M on the constructed rotor blade model and the radar at time t;

[0022] R' 0 =R 0 +l M (x,y,z).

[0023] Furthermore, the peak time shift Δt is calculated by the following formula:

[0024]

[0025] Furthermore, in step 2, the rotor single blade echo signal simulation is specifically:

[0026] As the width of the rotor blade increases, the phase of the corresponding rotation center O The larger the value, the smaller the rotor rotation frequency. The larger the rotor, the longer the time the full width of the rotor is used. By changing the rotor width, echo signal simulation with different aspect ratios is achieved to verify the contribution of the rotor blade width to the micro-Doppler echo.

[0027] Furthermore, in step 3, the rotor blade width estimation and error analysis based on the frequency peak time shift includes the following steps:

[0028] Step 3.1: Estimation of rotor blade width; the rotor length corresponds to the maximum micro-Doppler frequency shift of f L-max , the rotor blade length is obtained from formula (2):

[0029]

[0030] Where: L represents the length of the rotor blade;

[0031] The length-to-width ratio of the rotor blades is selected to be no less than 3:1. Within this range, the approximate width of the rotor blades is Combined with formula (4), the rotor blade width W is obtained:

[0032]

[0033] Step 3.2: Rotor blade width error analysis; The estimated relative error of the rotor blade width fluctuates around 1.09%, and the maximum relative error is less than 2%.

[0034] Further, in step 4, the rotor blade aspect ratio estimation and error analysis based on the frequency peak time shift are specifically performed as follows:

[0035] Step 41: Estimation of the rotor blade aspect ratio; the estimated value of the rotor blade width is related to the pitch angle. Combining formula (4) and formula (5), we get:

[0036]

[0037] From formula (6), we know that the expression of the aspect ratio of the rotor blade only contains the rotation speed ω and the peak time shift Δt, and has nothing to do with the pitch angle parameter;

[0038] Step 42: Rotor blade aspect ratio error analysis; the maximum relative error of rotor width identification is 3.33%, and the maximum relative error of aspect ratio is 4.52%.

[0039] The beneficial effects of the present invention compared with the prior art are as follows: based on the laser micro-Doppler effect, the present invention proposes a method for estimating the geometric parameters of UAV blades based on time-frequency peak time shift, which can effectively estimate the width and aspect ratio of the rotorcraft UAV, provide a theoretical basis for the high-accuracy identification of rotorcraft UAVs, and provide technical support for the high-precision identification of rotorcraft UAVs. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is the relative position of the rotor and the laser line of sight Figure 1 .

[0041] Figure 2 It is a time-frequency diagram of the rotor width with length:width:height=40:40:1.

[0042] Figure 3 It is a flow chart of the blade width extraction algorithm based on the time-shift method of the present invention.

[0043] Figure 4 It is a flow chart of the rotor aspect ratio estimation algorithm based on the time shift method of the present invention.

[0044] Figure 5 It is a simulation width estimation error diagram of the present invention.

[0045] Figure 6 It is a simulation aspect ratio estimation error diagram of the present invention.

[0046] Figure 7 It is the time-frequency diagram obtained by the experiment of the present invention.

[0047] Figure 8 is the relative position of the rotor and the laser line of sight Figure 2 . DETAILED DESCRIPTION

[0048] Specific implementation method 1: This implementation method discloses a method for estimating geometric parameters of UAV blades based on time-frequency peak time shift, and the method comprises the following steps:

[0049] Step 1: Construction of the multi-point scattering single blade echo model of the rotorcraft UAV;

[0050] Step 2: Simulation of the echo signal of a single rotor blade;

[0051] Step 3: Estimation and error analysis of rotor blade width based on peak frequency time shift;

[0052] Step 4: Estimation and error analysis of rotor blade aspect ratio based on frequency peak time shift.

[0053] Furthermore, the parameter estimation of the width and aspect ratio of the UAV rotor blades is experimentally verified (to ensure the effectiveness of the method in practical applications).

[0054] Furthermore, in step 1, the construction of the multi-point scattering single blade echo model of the rotor UAV is specifically as follows:

[0055] like Figure 8 As shown, the rotor blade rotation center is established and located at the coordinate origin O, the azimuth angle α = 0°, the pitch angle is β, and the distance from the laser radar to the rotor center is R 0 , the rotor blade height is z 0 , the distance from the scattering point on the rotor blade to the rotor center is l M ,R 0 >>l M , the scattering point rotates around the Z axis at a constant rotation speed ω; let the initial phase of point M on the rotor blade be The rotor blade reaches point M' after time t, and its phase is

[0056] From the single-point modeling, we know that the baseband signal returned by the scattering point M' at a distance R is:

[0057]

[0058] Where: f 0 represents the carrier frequency of the laser radar, λ represents the wavelength, and R M (t) represents the distance between point M on the constructed rotor blade model and the radar at time t, and σ represents the scattering rate;

[0059] The phase term is derived to obtain the micro-Doppler frequency shift caused by the rotation of point M on the rotor blade, that is:

[0060]

[0061] in: represents phase, t represents time, λ represents wavelength, and Δt represents peak time shift;

[0062] Rotor rotation During this period, the micro-Doppler frequency shift first increases from the maximum value f L-max It starts to decrease gradually, and then gradually increases to f L-max , that is, two peaks appear; the time used to generate the two peaks is the phase change The time used, this time difference is called the peak time shift Δt in the present invention;

[0063] The rotor blade is equivalent to a cuboid model with thickness, and the cuboid model is meshed. The points on the mesh are called scattering points, which replace the center point of the rotor blade scattering surface element. Let the coordinates of each scattering point be (x i ,y j ,z k ), where i, j, k are integers; the distance from the rotor center O to each scattering point is l M (x,y,z)=(x 2 +y 2 +z 2 ) 1 / 2 , the rotor blade echo signal is:

[0064]

[0065] Where: σ represents the scattering rate, f 0 represents the carrier frequency of the laser radar, R M (t) represents the distance between point M on the constructed rotor blade model and the radar at time t;

[0066] R' 0 =R 0 +l M (x,y,z).

[0067] Furthermore, the peak time shift Δt is calculated by the following formula:

[0068]

[0069] Furthermore, in step 2, the rotor single blade echo signal simulation is specifically:

[0070] like Figure 1 As shown in (see Figure 1 The larger the width of the rotor blade, the greater the phase of the corresponding rotation center O. The larger the value, the smaller the rotor rotation frequency. The larger the rotor, the longer the time the full width of the rotor is used. By changing the rotor width, echo signal simulation with different aspect ratios is achieved to verify the contribution of the rotor blade width to the micro-Doppler echo.

[0071] Furthermore, in step 3, the rotor blade width estimation and error analysis based on the frequency peak time shift includes the following steps:

[0072] Step 3.1: Estimation of rotor blade width; the rotor length corresponds to the maximum micro-Doppler frequency shift of f L-max , the rotor blade length is obtained from formula (2):

[0073]

[0074] Where: L represents the length of the rotor blade;

[0075] The length-to-width ratio of the rotor blades is selected to be no less than 3:1. Within this range, the approximate width of the rotor blades is Combined with formula (4), the rotor blade width W is obtained:

[0076]

[0077] Since the thickness of the rotor blades does not contribute to the echo signal waveform, the influence of the thickness of the rotor blades is not studied. In the simulation calculation, the thickness of the rotor blades is set to 1 / 40 of the length, and the width of the rotor blades is changed by adjusting the aspect ratio. The rotation speed ω, peak time shift Δt, and micro-Doppler maximum frequency shift f are extracted from the time-frequency diagram. L-max The extraction algorithm flow of rotor blade width is as follows: Figure 3 The rotation speed ω is obtained by two-dimensional Fourier transform, and the peak time shift Δt and the micro-Doppler maximum frequency shift f are obtained by edge extraction algorithm. L-max .

[0078] Step 3.2: Rotor blade width error analysis; The estimated relative error of the rotor blade width fluctuates around 1.09%, and the maximum relative error is less than 2%.

[0079] Further, in step 4, the rotor blade aspect ratio estimation and error analysis based on the frequency peak time shift are specifically performed as follows:

[0080] Step 41: Estimation of rotor blade aspect ratio; The estimated value of rotor blade width is related to the pitch angle. However, in reality, the pitch angle is generally unknown, and it is difficult to directly estimate the width. Combining formula (4) and formula (5), we get:

[0081]

[0082] From formula (6), we know that the expression of the aspect ratio of the rotor blade only contains the rotation speed ω and the peak time shift Δt, and has nothing to do with the pitch angle parameter (that is, the aspect ratio of the rotor blade can also be solved when the pitch angle is unknown). Figure 4The simulation calculation is performed according to the process shown in the figure. First, the rotor echo signal is obtained according to the established model. Then, the echo signal is converted into the original time-frequency diagram through short-time Fourier transform (STFT). Gaussian filtering and smoothing (i.e., image preprocessing) are performed to obtain the processed time-frequency diagram. Then, edge features are extracted from the time-frequency diagram through edge detection to obtain the peak time shift Δt and rotation speed ω. Finally, the rotor blade aspect ratio is calculated by formula (6).

[0083] Step 42: Rotor blade aspect ratio error analysis; the maximum relative error of rotor width identification is 3.33%, and the maximum relative error of aspect ratio is 4.52%.

[0084] Furthermore, in step 5, the parameter estimation experiment of the width and aspect ratio of the UAV rotor blade is verified; the specific process is:

[0085] The laser beam emitted by the laser is divided into two parts by the fiber optic beam splitter. One part is irradiated to the rotorcraft target by the transmitting optical system and then reflected. The reflected echo signal is transmitted to the receiving optical system and interferes with the other part of the light in the fiber optic beam combiner. The signal is transmitted to the balanced detector to convert the optical signal into an electrical signal, which is collected by the acquisition card and finally processed by the computer to obtain the time-frequency diagram (see Figure 7 ).

[0086] The experiment uses a pulsed laser radar with a laser wavelength of 1064nm (to ensure the measurement accuracy of the laser radar system), the distance between the laser and the target is 80m, the spot diameter of the laser beam is 3cm, the motor rotation speed is 500r / min, and the target rotation power supply voltage is 6V.

[0087] Embodiment 1:

[0088] This embodiment discloses a method for estimating geometric parameters of UAV blades based on time-frequency peak time shift, the method comprising the following steps:

[0089] Step 1: Derivation of the rotor time-frequency peak time shift; the relative position of the rotor and the laser radar is as follows: Figure 1 As shown in the figure, the distance between the laser radar and the target rotor is much greater than the rotor length L, and the laser can be seen as a parallel light (such as the parallel line with an arrow on the right side of the figure). To simplify the calculation, the laser radar is equivalent to point Q, and the straight line QO is parallel to the laser radar line of sight. The distance from the laser radar Q to the rotation center O of the rotor blade is R0, and the distance from Q to the top point M of the rotor blade is RM. θ is the angle from the straight line OM to OM′, so:

[0090]

[0091] Where: f M represents micro-Doppler shift;

[0092] The micro-Doppler frequency shift generated by each scattering point on the rotor blade is related to ω and θ. The rotation speed of the rotor is constant. When n is a positive integer, n=1,2,3,....;

[0093] The maximum micro-Doppler frequency shift band affected by the rotor blade length is:

[0094]

[0095] When θ=±nπ, the micro-Doppler frequency shift reaches the minimum value of zero, so the time-frequency diagram changes periodically. Figure 1 As shown in the small figure in the upper right corner, points A and B are the two vertices of the blade, and the width of the rotor blade is AB. When the phase change rotates to the point where OB is perpendicular to OQ, the micro-Doppler frequency shift caused by the blade length reaches its maximum value, and the blade length is approximately equal to the rotation radius. During this period, the micro-Doppler frequency shift first increases from the maximum value f L-max It starts to decrease gradually, and then gradually increases to f L-max , that is, two peaks appear, and the time taken to generate the two peaks is the phase change The time used, this time difference is called the peak time shift Δt in the present invention:

[0096]

[0097] Step 2: Table 1 gives the parameters used in the simulation;

[0098] Table 1 Rotor single blade echo signal simulation parameters

[0099]

[0100] Depend on Figure 2 It can be seen that the simulation time-frequency diagram shows periodic changes, with 4 cycles per second, corresponding to the rotor rotation frequency of 4Hz. Figure 2 It can be clearly seen in the dotted box that there are two peaks, T represents the time when the peak is generated, and F represents the frequency corresponding to the peak. From the laser frequency, it can be seen that the maximum frequency shift generated by the echo of a single rotor blade at time T is 11MHz. The interval between the two peaks is the time shift. When the aspect ratio decreases in proportion, the time shift decreases as the rotor width decreases. When the rotation frequency is the same, the longer the time it takes for the peak to be formed from the vertex at one end of the width to the other vertex, the greater the interval between the two peaks and the greater the time shift. The above simulation verification and analysis show that the time shift effect does exist, and the time shift method can be used to estimate the width of the rotor blade.

[0101] Step 3: If Figure 3As shown in the figure (the flow chart of blade width extraction algorithm), it is assumed that the pitch angle β of the rotor is 30° and the rotor speed f 0 =4Hz, the rotor length corresponds to the maximum micro-Doppler frequency shift of f L-max , rotation speed ω=8πrad / s, laser wavelength λ=1064nm;

[0102] The width estimation error is Figure 5 As shown by Figure 5 It can be seen that the relative error fluctuates around 1.09%, and the maximum relative error is less than 2%. In summary, the time shift method proposed in the present invention can estimate the rotor width well, and the estimation result is highly reliable.

[0103] Step 4: The absolute error and relative error of aspect ratio estimation are as follows Figure 6 As shown, the relative error fluctuates between 0 and 2%. Similarly, the relative error shows an upward trend with the increase of the aspect ratio, and the actual maximum relative error is only 1.6%. Similar to the width estimation, the aspect ratio estimation error mainly comes from the discrete error introduced by the model gridding and the time shift extraction error, while the frequency shift extraction result does not affect the aspect ratio estimation. Therefore, the relative error of the aspect ratio estimation result is smaller than the relative error of the width estimation result. From the above analysis, it can be seen that the time shift method proposed in the present invention has a good estimation effect on the rotor aspect ratio parameters and is highly reliable.

[0104] Step 5: Experimental verification of the estimation of the width and aspect ratio parameters of the UAV rotor blades. The length of a single rotor blade is 9 mm, and its rotation frequency is 1.59 Hz. The two vertical lines in the time-frequency graph are the peaks formed by the two tip points of the rotor width, as shown in Figure 5. Figure 7 As shown. For each set of data, 5 sets are taken to calculate the average value, and the time when the time shift peak occurs and the corresponding maximum frequency shift are recorded. The experimental and simulation results are as follows:

[0105] Table 2 Comparison of experimental results and simulation results

[0106]

[0107] Since the accuracy of the simulation is higher than that of the experiment, it can be seen from Table 2 that the maximum relative error of the rotor width identification is 3.33%, and the maximum relative error of the aspect ratio is 4.52%. The above comparison results prove the feasibility of the UAV blade geometric parameter estimation method proposed in the present invention, which can extract the rotor width and aspect ratio with high accuracy.

[0108] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other forms of assembly without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations within the meaning and range of equivalents of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for estimating geometric parameters of UAV blades based on time-frequency peak time shift, characterized by: The method comprises the following steps: Step 1: Construction of the multi-point scattering single blade echo model of the rotorcraft UAV; Step 2: Simulation of the echo signal of a single rotor blade; Step 3: Estimation and error analysis of rotor blade width based on peak frequency time shift; Step 4: Estimation and error analysis of rotor blade aspect ratio based on frequency peak time shift.

2. The method for estimating geometric parameters of UAV blades based on time-frequency peak time shift according to claim 1 is characterized in that: In step 1, the construction of the multi-point scattering single blade echo model of the rotor UAV is specifically as follows: The rotor blade rotation center is established and located at the coordinate origin O, the azimuth angle α = 0°, the pitch angle is β, the distance from the laser radar to the rotor center is R0, the rotor blade height is z0, and the distance from the scattering point on the rotor blade to the rotor center is l M ,R0>>l M , the scattering point rotates around the Z axis at a constant rotation speed ω; let the initial phase of point M on the rotor blade be The rotor blade reaches point M' after time t, and its phase is From the single-point modeling, we know that the baseband signal returned by the scattering point M' at a distance R is: Where: f0 represents the carrier frequency of the laser radar, λ represents the wavelength, R M (t) represents the distance between point M on the constructed rotor blade model and the radar at time t, and σ represents the scattering rate; The phase term is derived to obtain the micro-Doppler frequency shift caused by the rotation of point M on the rotor blade, that is: in: represents phase, t represents time, λ represents wavelength, and Δt represents peak time shift; The rotor blade is equivalent to a cuboid model with thickness, and the cuboid model is meshed. The points on the mesh are called scattering points, which replace the center point of the rotor blade scattering surface element. Let the coordinates of each scattering point be (x i ,y j ,z k ), where i, j, k are integers; the distance from the rotor center O to each scattering point is l M (x,y,z)=(x 2 +y 2 +z 2 ) 1 / 2 , the rotor blade echo signal is: Where: σ represents the scattering rate, f0 represents the carrier frequency of the laser radar, R M (t) represents the distance between point M on the constructed rotor blade model and the radar at time t; R'0=R0+l M (x,y,z)。 3. The method for estimating geometric parameters of UAV blades based on time-frequency peak time shift according to claim 2 is characterized in that: The peak time shift Δt is calculated from the following formula:

4. The method for estimating geometric parameters of UAV blades based on time-frequency peak time shift according to claim 1, characterized in that: In step 2, the rotor single blade echo signal simulation is specifically as follows: As the width of the rotor blade increases, the phase of the corresponding rotation center O The larger the value, the smaller the rotor rotation frequency. The larger the rotor, the longer the time the full width of the rotor is used. By changing the rotor width, echo signal simulation with different aspect ratios is achieved to verify the contribution of the rotor blade width to the micro-Doppler echo.

5. The method for estimating geometric parameters of UAV blades based on time-frequency peak time shift according to claim 1, characterized in that: In step 3, the rotor blade width estimation and error analysis based on the peak frequency time shift includes the following steps: Step 3.1: Estimation of rotor blade width; the rotor length corresponds to the maximum micro-Doppler frequency shift of f L-max , the rotor blade length is obtained from formula (2): Where: L represents the length of the rotor blade; The length-to-width ratio of the rotor blades is selected to be no less than 3:

1. Within this range, the approximate width of the rotor blades is Combined with formula (4), the rotor blade width W is obtained: Step 3.2: Rotor blade width error analysis; The estimated relative error of the rotor blade width fluctuates around 1.09%, and the maximum relative error is less than 2%.

6. The method for estimating geometric parameters of UAV blades based on time-frequency peak time shift according to claim 5, characterized in that: In step 4, the rotor blade aspect ratio estimation and error analysis based on the peak frequency time shift are performed; the specific steps are as follows: Step 41: Estimation of the rotor blade aspect ratio; the estimated value of the rotor blade width is related to the pitch angle. Combining formula (4) and formula (5), we get: From formula (6), we know that the expression of the aspect ratio of the rotor blade only contains the rotation speed ω and the peak time shift Δt, and has nothing to do with the pitch angle parameter; Step 42: Rotor blade aspect ratio error analysis; the maximum relative error of rotor width identification is 3.33%, and the maximum relative error of aspect ratio is 4.52%.

Images