A multi-observation based spatial neighbor extended target angle estimation method

Abstract

The estimation of spatially proximal extended target angle is always a hot and difficult research topic. The present invention proposes a method for estimating the angle of spatially proximal extended target based on multi-observation. According to the scattering characteristics of the extended target between adjacent pulses, the motion state of the extended target is predicted, then the maximum likelihood estimation method is used to estimate the angle of the spatially proximal extended target, and the optimization process of the RELAX algorithm is given.

Description

Technical Field

[0001] This invention belongs to the field of broadband target angle estimation technology, and relates to a method for estimating the angle of a spatially proximate extended target based on a multi-observation spatial proximity extended target angle estimation method. Background Technology

[0002] High-precision target angle estimation is crucial for target tracking. In scenarios such as aircraft formation flying, when multiple targets are close together in space, they are often referred to as spatially proximate targets. Due to errors in target parameter measurement, target tracks may intersect when tracking spatially proximate targets. Furthermore, as radar range resolution increases, it gradually becomes smaller than the target size, causing the target to expand into multiple scattering points. In this case, spatially proximate extended targets may have multiple scattering points located within the same range cell, posing a challenge to angle estimation for these targets.

[0003] Monopulse radar is a radar angle measurement technique that obtains precise target position information by comparing the received signals of two or more antenna beams simultaneously. In practical applications, monopulse radar is widely used due to its advantages such as high angle measurement accuracy and ease of engineering implementation. Methods for estimating the angle of spatially proximal targets include complex monopulse ratio, maximum likelihood estimation, and spectral estimation. However, these methods are mainly discussed under narrowband radar conditions, and there is currently no literature discussing them under broadband conditions. Summary of the Invention

[0004] Estimating the angle of spatially proximate targets has always been a research hotspot and a challenge. This invention proposes a multi-observation method for estimating the angle of spatially proximate extended targets. This method predicts the motion state of the extended target based on the scattering characteristics between adjacent pulses, then uses the maximum likelihood estimation method to estimate the angle of the spatially proximate extended target, and provides the optimization process of the RELAX algorithm.

[0005] Assuming the radar transmits a linear frequency modulated signal, the flowchart of a multi-observation spatial proximity extended target angle estimation method is as follows: Figure 1 As shown, the specific steps include the following:

[0006] Step 1: Receive the first m Each sum and difference channel echo pulse;

[0007] Step 2, for the first m Matched filtering is performed on the echo pulses of the sum and difference channels to obtain the first... m One-dimensional distance image of the sum and difference channels at each pulse moment and :

[0008] (1)

[0009] where there are I extended targets in the space, is the sum channel gain of the th scatterer of the th extended target under the th pulse, is the amplitude of the th scatterer of the B th extended target under the c th pulse, f is the bandwidth, is the speed of light, c is the carrier frequency, is the complex Gaussian white noise, is the distance from the th scatterer of the

[0010] th extended target to the radar, which can be written as:

[0011] where , and are the distance, velocity and acceleration of the th scatterer of the th extended target at the th pulse time.

[0012] (1) can be written in vector form as:

[0013] (3)

[0014] where

[0015] (4)

[0016] (5)

[0017] Similarly, the echo signal of the difference channel can also be written as:

[0018] (6)

[0019] (7)

[0020] (8)

[0021] Step three, extended target echo state prediction. The observation of the extended target motion process usually needs multiple pulses, assuming that the initial motion state of the extended target is known, i.e. the angle of the extended target in (5) and its corresponding state It is known that the scattering characteristics of the extended target remain essentially unchanged between adjacent pulses. Based on this assumption, we only need to estimate the velocity of the extended target at each pulse moment to predict the echo state of the two extended targets at the next pulse moment, that is:

[0022] (9)

[0023] in, It is the state transition matrix. It is the first The state at each pulse moment.

[0024] According to the above method, the first Predicting the echo state of each pulse moment and channel as follows: Figure 2 As shown, it can be written as a formula:

[0025] (10)

[0026] in, , It is the length of the sliding window. yes The window length of the captured extended target is The sum and difference channel echoes.

[0027] Step 4: Estimate the amplitude of the extended target and difference channels using the RELAX method. The flowchart of its angle measurement principle is as follows: Figure 3 As shown.

[0028] Due to (3) neutralizing channel noise It's complex Gaussian white noise, so it's related to the channel. The probability density function can be written as:

[0029] (11)

[0030] in, It is the echo length of the channel.

[0031] Available and channel amplitude Maximum likelihood estimation:

[0032] (12)

[0033] Similarly, the difference channel amplitude can be obtained. Maximum likelihood estimation:

[0034] (13)

[0035] According to (12) and (13), to obtain the amplitude of the extended target and difference channels, it is necessary to first reconstruct the echo states of the two extended targets. and .and and The reconstruction and the speed of expanding the target at each pulse moment Regarding (assuming the extended target moves at a constant speed). It should be noted that, due to the higher gain of the signal in the same channel compared to the difference channel, and also to reduce the computational load of the algorithm, the extended target velocity estimated by the echo signal in the same channel is used instead of the extended target velocity in the difference channel.

[0036] According to the least squares estimation, we can obtain... The estimate is:

[0037] (14)

[0038] in, It is obtained by the method in step two.

[0039] The RELAX algorithm, as a fast parameter estimation method, can estimate the parameters of each extended target from the echo sequentially according to the target amplitude. Therefore, this section will take two extended targets as examples to introduce how to use the RELAX algorithm to optimize (12)~(14).

[0040] When estimating the first When expanding the target, we have:

[0041] (15)

[0042] (16)

[0043] in, It is the estimated first The extended target speed. Based on (14) and (15), the first... The speed of each extended target:

[0044] (17)

[0045] Will Substituting (12) and (13) into the equation, we can obtain the first... The amplitude of each extended target and difference channel:

[0046] (18)

[0047] (19)

[0048] The steps of the RELAX optimization algorithm are given below:

[0049] Step 1: Initialization

[0050] A Let the parameters ;

[0051] B Obtain from (15) and (16) and , and update the parameters according to (17)-(19) .

[0052] Step 2: Iterative estimation

[0053] A Let the parameters ;

[0054] B Obtain from (15) and (16) and , and update the parameters according to (17)-(19) ;

[0055] C Obtain from and (15), (16) and , and update the parameters according to (17)-(19) ;

[0056] D Repeat steps B and C until the iteration stopping condition is met.

[0057] wherein the cost function can be set as:

[0058] (20)

[0059] wherein, When the iteration termination condition is met, the sum-difference channel amplitudes of the two extended targets are output.

[0060] Step five, obtain the angle of the first extended target according to the monopulse ratio:

[0061] (21)

[0062] wherein, is the angle error slope. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 is the flow chart for spatially adjacent extended target estimation;

[0064] Figure 2 is the flow chart for extended target state prediction;

[0065] Figure 3 is the principle diagram of the RELAX algorithm; ​

[0066] Figure 4 For the first simulation scenario, the RMSE results of the proposed method and the traditional wideband angle measurement method are shown in Table 1.

[0067] Figure 5 For the first simulation scenario, the RMSE results of the proposed method and the traditional wideband angle measurement method are shown in Table 1.

[0068] Figure 6 For the second simulation scenario, the RMSE results of the proposed method and the traditional wideband angle measurement method are shown in Table 2. DETAILED DESCRIPTION

[0069] An experimental simulation applying the present application is given below, and the specific process is described.

[0070] The radar parameters are shown as follows:

[0071]

[0072] The spatially adjacent extended target motion scenario is shown in Fig. 1, where the velocities of the two extended targets are 50 and 60 m / s, respectively, and the signal-to-noise ratio of the first extended target is 3 dB higher than that of the second extended target. Figure 4 The angle tracking estimation results of the two extended targets under the two simulation scenarios are shown in Figs. 2 and 3, respectively.

[0073] The angle tracking estimation results of the two extended targets under the two simulation scenarios are shown in Figs. 2 and 3, respectively. Figure 5 Figure 6 The experimental simulation results show that the RMSE of the angle estimation of the proposed method is smaller than that of the traditional wideband angle measurement method, and the angle estimation performance is better, especially when the two extended targets overlap each other in the one-dimensional range image (see the dashed box part in the figure), compared with the traditional wideband angle measurement method, the proposed method can effectively estimate the angles of the two extended targets.​​​​

Claims

1. A method for spatially proximal extended target angle estimation based on multiple observations, characterized in that, The method comprises the following steps: Step one, receiving the first m and difference channel echo pulses; Step 2, for the first m Matched filtering is performed on the echo pulses of the sum and difference channels to obtain the first... m One-dimensional distance image of the sum and difference channels at each pulse moment and ; Step three, extending the target echo state prediction; Step four, estimate the spread target and difference channel amplitudes using the RELAX method; and channel noise is a complex Gaussian white noise, and channel The probability density function of is ； wherein is the echo length of the channel; and channel amplitude The maximum likelihood estimate of the channel amplitude is: ； difference channel amplitudes The maximum likelihood estimate of the difference channel amplitudes ; Step five, obtaining the extended target angle according to the single pulse ratio.

2. The multi-observation based spatially-adjacent extended target angle estimation method of claim 1, wherein, Step two matches the sum and difference channel echoes to obtain the first m pulsed time and the difference channel one-dimensional range image m at the first pulsed time and : ； wherein it is assumed that there are I extended targets in the space, is the sum channel gain of the th extended target under the th pulse, is the amplitude of the th scattering point of the th extended target, B is the bandwidth, c is the speed of light, is the carrier frequency, is the complex Gaussian white noise, is the distance of the th scattering point of the th extended target to the radar.

3. The multi-observation based spatially-adjacent extended target angle estimation method of claim 1, wherein, Step 3 The prediction of the echo state of the channel at each pulse moment can be expressed by the formula: ； wherein, , is the sliding window length, is the window length of the intercepted extended target is sum and difference channel echoes.

4. The multi-observation based spatially-adjacent extended target angle estimation method of claim 1, wherein, The angle of the extended target is obtained in step five by neutralizing the difference channel amplitude ​ ； wherein is the angular error slope.

Images