A non-matched filtering method for radar frequency-phase composite coded signals

Abstract

A method for unmatched filtering of radar frequency-phase composite coded signals is disclosed, comprising: S1. Generation of frequency-phase composite coded signals; S2. Design and modeling of unmatched filters; S3. Optimization design of unmatched filters; S4. Sub-pulse unmatched filtering; S5. Sub-pulse non-coherent accumulation; ultimately obtaining a low sidelobe pulse compression result from the unmatched filtering of frequency-phase composite coded signals. Compared with direct matched filtering and non-coherent accumulation methods for sub-pulse matched filtering of frequency-phase composite coded signals, the unmatched filtering and non-coherent accumulation methods of this invention result in a lower sidelobe level. This invention allows for offline design of frequency-phase composite coded signals and unmatched filters with low pulse compression performance, and has broad application prospects in air defense radar, anti-missile radar, and inverse synthetic aperture radar.

Description

Technical Field

[0001] The invention belongs to the field of pulse compression radar systems, and in particular relates to a non-matching filtering method for radar frequency-phase composite coded signals. Background Art

[0002] The frequency-phase composite coded signal pulse contains multiple sub-pulse signals, each sub-pulse signal is modulated in phase coding, and different carrier frequencies are used between each sub-pulse. The use of complex modulation makes it difficult for the signal to be intercepted by reconnaissance, and also increases the difficulty of local estimation of signal parameters and structure, which is of great significance to improving the battlefield survivability of radar. In addition, sub-pulse processing of the frequency-phase composite coded signal and elimination of the interfered sub-pulse signal can improve the radar's anti-narrowband aiming interference performance.

[0003] In frequency-composite phase-coded signals, sub-pulses usually use the longest linear shift register sequence, and the processing methods include direct matched filtering and sub-pulse matched filtering pulse compression. However, the pulse compression results obtained by matched filtering processing all have the problem of too high pulse compression sidelobe level, which causes the sidelobe of strong scattering targets to mask weak scattering targets in multi-target detection environments.

[0004] In view of the above reasons, a non-matched filtering method for radar frequency phase complex coded signals is developed. Summary of the invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art, provide a non-matched filtering method for radar frequency-phase composite coded signals, establish a non-matched filter design optimization model for pulse compression sidelobe suppression based on non-coherent accumulation processing for frequency-phase composite coded signals, optimize and design the non-matched filter through penalty function and quasi-Newton gradient descent method, finally obtain the pulse compression result through non-matched filtering and non-coherent accumulation processing, and finally suppress the pulse compression sidelobe level of the frequency-phase composite coded signal. Compared with the direct matched filtering and sub-pulse matched filtering non-coherent accumulation processing methods of the frequency-phase composite coded signals, the non-matched filtering and non-coherent accumulation processing of the present invention has a lower sidelobe level; the present invention can design frequency-phase composite coded signals and non-matched filters with low pulse compression performance offline, and has broad application prospects in the fields of air defense radar, anti-missile radar and inverse synthetic aperture radar.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a non-matching filtering processing method for radar frequency-phase composite coded signals, s1. Generate the longest linear shift register sequence according to a specific feedback coefficient, and perform composite coding on multiple longest linear shift register sequences according to the frequency coding sequence to generate a frequency-phase composite coded signal; s2. For each sub-pulse signal in the frequency-phase composite coded signal, a pulse compression mathematical expression is established based on non-matching filtering pulse compression processing, and on this basis, based on the principle of sub-pulse non-coherent accumulation processing, a low pulse compression sidelobe objective function of non-matching filtering and non-coherent accumulation processing of the frequency-phase composite coded signal is established. , and establish a non-matching filter design model under the signal processing gain loss constraint condition; s3. For the non-matching filter design mathematical model, the constrained optimization problem is converted into an unconstrained optimization problem based on the penalty function principle, and the quasi-Newton gradient descent method is used to solve it to obtain the optimized non-matching filter; s4. Using the optimized non-matching filter, each sub-pulse in the frequency-phase composite coded signal is subjected to non-matching filtering pulse compression processing to obtain the sub-pulse non-matching filtering result; s5. For the non-matching filtering result of each sub-pulse in the frequency-phase composite coded signal, a non-coherent accumulation process is performed, and finally a pulse compression result with low sidelobe characteristics is obtained.

[0007] Furthermore, the specific steps of the non-matched filtering processing method of the radar frequency phase composite coded signal are as follows:

[0008] Assume that the number of radar sub-pulses is M and the bandwidth of each sub-pulse signal is B r , time width is T r The intra-pulse modulation pattern is the longest linear shift register sequence, and the sub-pulse carrier frequencies are f m ;

[0009] Step 1: Generate frequency-phase composite coded signals based on radar parameter design. The mth sub-pulse phase coded signal is:

[0010]

[0011] Where N represents the number of code elements, represents the symbol phase, and the longest linear shift register sequence is ±π, w(τ) represents the pulse width T r The rectangular pulse signal, w sub (τ-nT sub ) indicates that the pulse width is T sub A rectangular pulse function, and T sub =1 / B r On this basis, the frequency-phase composite coded signal can be obtained as

[0012]

[0013] Step 2: For each sub-pulse signal of the frequency composite phase coded signal, a discrete sampling process is performed. Each sub-pulse code element samples a point, and the sub-pulse discrete signal is obtained as follows:

[0014]

[0015] The unmatched filter of each sub-pulse discrete signal is defined as:

[0016] h m =[h m1 h m2 … h mN ] T (4)

[0017] Then the sub-pulse discrete signal non-matched filtering output is:

[0018]

[0019] Then the sub-pulse non-coherent accumulation output is:

[0020]

[0021] where |r m | represents the modulus value of the vector element by element, r represents a vector of length 2N-1, and the Nth element is the pulse compression peak value. The sub-pulse non-coherent accumulation output vector is defined as:

[0022]

[0023] Then the cost function of peak sidelobe suppression can be approximated as:

[0024]

[0025] Among them, ||s|| p represents the p-norm of the vector. In order to control the signal processing gain loss, the following constraints are introduced for each sub-pulse unmatched filtering:

[0026]

[0027] Among them, b max represents the expected output peak value of the sub-pulse non-matching filter. In summary, the non-matching filter design problem for frequency-complex phase-coded signals is:

[0028] min f(h1,h2,…,h M )

[0029]

[0030]

[0031] Step 3: For the non-matched filter design problem, use the penalty function method to transform it into an unconstrained problem:

[0032]

[0033] In this unconstrained optimization problem, each sub-objective function is second-order differentiable and can be transformed into a real function of the real and imaginary parts of the non-matched filter, so it can be solved using the quasi-Newton gradient descent method;

[0034] Step 4: After the sub-pulse signal of the frequency-phase composite coded signal is processed by the non-matched filter optimized and designed in the third step, the non-matched filtering pulse compression result of each sub-pulse signal can be obtained;

[0035] Step 5: For the non-matched filtering pulse compression results of each sub-pulse, non-coherent accumulation is performed using equation (6) to obtain the final radar pulse compression result after sidelobe suppression, and subsequent target detection and parameter estimation are performed.

[0036] The beneficial effects of the present invention are as follows: the present invention establishes a non-matching filter design optimization model for pulse compression sidelobe suppression based on non-coherent accumulation processing for frequency-phase composite coded signals, optimizes and designs non-matching filters through penalty functions and quasi-Newton gradient descent methods, and finally obtains pulse compression results through non-matching filtering and non-coherent accumulation processing, thereby finally suppressing the pulse compression sidelobe level of the frequency-phase composite coded signal. Compared with the direct matching filtering and sub-pulse matching filtering non-coherent accumulation processing methods for frequency-phase composite coded signals, the non-matching filtering and non-coherent accumulation processing of the present invention obtains a lower sidelobe level; the present invention can design frequency-phase composite coded signals and non-matching filters with low pulse compression performance offline, and has broad application prospects in the fields of air defense radar, anti-missile radar, and inverse synthetic aperture radar; the parts not described in detail in the present invention are existing common technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention will be further described below in conjunction with the accompanying drawings:

[0038] Figure 1 It is a flow chart of the non-matched filtering processing method of radar frequency phase composite coded signal;

[0039] Figure 2 It is the time-frequency result of frequency-phase composite coded signal;

[0040] Figure 3 It is the result of direct matched filtering of frequency-phase composite coded signal;

[0041] Figure 4 Sub-pulse matched filtering and non-coherent accumulation results for frequency-phase composite coded signal;

[0042] Figure 5 To optimize the design of the real part of the non-matched filter for the 4 sub-pulse signals;

[0043] Figure 6 It is the result of non-matched filtering and non-coherent accumulation of sub-pulses of frequency-phase composite coded signal. DETAILED DESCRIPTION

[0044] The present invention is further described in detail below in conjunction with embodiments and specific implementation modes:

[0045] Example

[0046] Assume that the number of radar sub-pulses is M and the bandwidth of each sub-pulse signal is B r , time width is T r The intra-pulse modulation pattern is the longest linear shift register sequence, and the sub-pulse carrier frequencies are f m ;

[0047] Step 1: Generate frequency-phase composite coded signals based on radar parameter design. The mth sub-pulse phase coded signal is:

[0048]

[0049] Where N represents the number of code elements, represents the symbol phase, and the longest linear shift register sequence is ±π, w(τ) represents the pulse width T r The rectangular pulse signal, w sub (τ-nT sub ) indicates that the pulse width is T sub A rectangular pulse function, and T sub =1 / B r On this basis, the frequency-phase composite coded signal can be obtained as

[0050]

[0051] Step 2: For each sub-pulse signal of the frequency composite phase coded signal, a discrete sampling process is performed. Each sub-pulse code element samples a point, and the sub-pulse discrete signal is obtained as follows:

[0052]

[0053] The unmatched filter of each sub-pulse discrete signal is defined as:

[0054] h m =[h m1 h m2 … h mN ] T (4)

[0055] Then the sub-pulse discrete signal non-matched filtering output is:

[0056]

[0057] Then the sub-pulse non-coherent accumulation output is:

[0058]

[0059] Among them, |r m | represents the modulus value of the vector element by element, r represents a vector of length 2N-1, and the Nth element is the pulse compression peak value. The sub-pulse non-coherent accumulation output vector is defined as:

[0060]

[0061] Then the cost function of peak sidelobe suppression can be approximated as:

[0062]

[0063] Among them, ||s|| p represents the p-norm of the vector. In order to control the signal processing gain loss, the following constraints are introduced for each sub-pulse unmatched filtering:

[0064]

[0065] Among them, b max represents the expected output peak value of the sub-pulse non-matching filter. In summary, the non-matching filter design problem for frequency-complex phase-coded signals is:

[0066]

[0067] Step 3: For the non-matched filter design problem, use the penalty function method to transform it into an unconstrained problem:

[0068]

[0069] In this unconstrained optimization problem, each sub-objective function is second-order differentiable and can be transformed into a real function of the real and imaginary parts of the non-matched filter, so it can be solved using the quasi-Newton gradient descent method;

[0070] Step 4: After the sub-pulse signal of the frequency-phase composite coded signal is processed by the non-matched filter optimized and designed in the third step, the non-matched filtering pulse compression result of each sub-pulse signal can be obtained;

[0071] Step 5: For the non-matched filtering pulse compression results of each sub-pulse, non-coherent accumulation is performed using equation (6) to obtain the final radar pulse compression result after sidelobe suppression, and subsequent target detection and parameter estimation are performed.

[0072] Figure 2-Figure 6 The processing results of the simulation experiment using the specific implementation of the present invention are shown in Figure 1. Assume that the number of radar sub-pulses is 4, the bandwidth of each sub-pulse signal is 10 MHz, the duration is 12.7 us, the intra-pulse modulation pattern is the longest linear shift register sequence, and the sub-pulse carrier frequencies are 40 MHz, 20 MHz, 30 MHz and 10 MHz respectively.

[0073] Figure 2 It is the time-frequency result of the frequency-phase composite coded signal obtained by short-time Fourier transform, wherein the bandwidth of each sub-pulse signal is about 10 MHz, the time width is 12.7 us, and the sub-pulse carrier frequency is consistent with the expected setting parameters, indicating that the present invention can design a frequency-phase composite coded signal.

[0074] Figure 3 It is the pulse compression result obtained by directly matching the frequency-phase composite coded signal, and its peak sidelobe is -23.36dB.

[0075] Figure 4 It is the pulse compression result of sub-pulse matched filtering and non-coherent accumulation of frequency-phase composite coded signal, and its peak sidelobe is -25.82dB.

[0076] Figure 5 The real part result of the non-matched filter designed for the sub-pulse of the frequency-phase composite coded signal, where p = 10, the expected signal processing gain loss of each sub-pulse non-matched filter is -1dB, and its energy meets the constraint condition and It shows that the non-matching filter with the optimal design of the present invention can control the gain loss of the sub-pulse non-matching filter signal processing.

[0077] Figure 6 The pulse compression result obtained by non-matching filtering and non-coherent accumulation processing of frequency-phase composite coded signals has a pulse compression signal processing gain loss of -1dB and a peak sidelobe ratio of -30.28dB. Compared with the pulse compression results of direct matched filtering and sub-pulse matched filtering and non-coherent accumulation, the peak sidelobe ratio can be improved by 6.92dB and 4.46dB respectively by the method of the present invention. The results show that the present invention can improve the peak sidelobe performance of the frequency-phase composite coded signal with a relatively small signal processing gain loss, thereby helping to improve the radar's detection performance for weak targets.

[0078] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for processing radar frequency-phase composite coded signals by non-matching filtering, characterized in that: s1. Generate the longest linear shift register sequence according to a specific feedback coefficient, and generate a composite encoding of multiple longest linear shift register sequences according to the frequency coding sequence to generate a frequency phase composite coding signal; s2. For each sub-pulse signal in the frequency-phase composite coded signal, a pulse compression mathematical expression is established based on the non-matching filter pulse compression processing. On this basis, based on the principle of sub-pulse non-coherent accumulation processing, a low pulse compression sidelobe objective function of non-matching filtering and non-coherent accumulation processing of the frequency-phase composite coded signal is established, and a non-matching filter design model is established under the signal processing gain loss constraint condition; s3. For the non-matching filter design mathematical model, the constrained optimization problem is converted into an unconstrained optimization problem based on the penalty function principle, and the quasi-Newton gradient descent method is used to solve it to obtain an optimized non-matching filter; s4. Using the optimized non-matched filter, each sub-pulse in the frequency-phase composite coded signal is subjected to non-matched filtering pulse compression processing to obtain a sub-pulse non-matched filtering result; s5. Perform non-matched filtering on each sub-pulse in the frequency-phase composite coded signal, and perform non-coherent accumulation processing to finally obtain a pulse compression result with low sidelobe characteristics.

2. The method for processing radar frequency phase complex coded signals by non-matched filtering according to claim 1, characterized in that: The specific steps of the non-matched filtering processing method of the radar frequency phase composite coded signal are: Assume that the number of radar sub-pulses is M and the bandwidth of each sub-pulse signal is B r , time width is T r The intra-pulse modulation pattern is the longest linear shift register sequence, and the sub-pulse carrier frequencies are f m ; Step 1: Generate frequency-phase composite coded signals based on radar parameter design. The mth sub-pulse phase coded signal is: Where N represents the number of code elements, represents the symbol phase, and the longest linear shift register sequence is ±π, w(τ) represents the pulse width T r The rectangular pulse signal, w sub (τ-nT sub ) indicates that the pulse width is T sub A rectangular pulse function, and T sub =1 / B r On this basis, the frequency-phase composite coded signal can be obtained as Step 2: For each sub-pulse signal of the frequency composite phase coded signal, a discrete sampling process is performed. Each sub-pulse code element samples a point, and the sub-pulse discrete signal is obtained as follows: The unmatched filter of each sub-pulse discrete signal is defined as: h m =[h m1 h m2 …h mN ] T (4) Then the sub-pulse discrete signal non-matched filtering output is: Then the sub-pulse non-coherent accumulation output is: where |r m | represents the modulus value of the vector element by element, r represents a vector of length 2N-1, and the Nth element is the pulse compression peak value. The sub-pulse non-coherent accumulation output vector is defined as: Then the cost function of peak sidelobe suppression can be approximated as: Among them, ||s|| p represents the p-norm of the vector. In order to control the signal processing gain loss, the following constraints are introduced for each sub-pulse unmatched filtering: Among them, b max represents the expected output peak value of the sub-pulse non-matching filter. In summary, the non-matching filter design problem for frequency-complex phase-coded signals is: Step 3: For the non-matched filter design problem, use the penalty function method to transform it into an unconstrained problem: In this unconstrained optimization problem, each sub-objective function is second-order differentiable and can be transformed into a real function of the real and imaginary parts of the non-matched filter, so it can be solved using the quasi-Newton gradient descent method; Step 4: After the sub-pulse signal of the frequency-phase composite coded signal is processed by the non-matched filter optimized and designed in the third step, the non-matched filtering pulse compression result of each sub-pulse signal can be obtained; Step 5: For the non-matched filtering pulse compression results of each sub-pulse, non-coherent accumulation is performed using equation (6) to obtain the final radar pulse compression result after sidelobe suppression, and subsequent target detection and parameter estimation are performed.

Images