Radar communication integrated system

Abstract

The invention relates to a radar communication integrated system, and belongs to the technical field of radar communication signal integration, and the system comprises an integrated signal transmitting end which is used for obtaining an N-bit information stream; determining a target initial frequency mapped by the first N1 pieces of bit information in the N-bit information stream, and determining a target frequency modulated frequency mapped by the last N2 pieces of bit information in the N-bit information stream; wherein the sum of N1 and N2 is equal to N, and N1, N2 and N are all positive integers; and generating and transmitting a target LFM radar signal according to the target initial frequency and the target frequency modulation rate. According to the invention, the constant enveloping property and phase continuity of the transmitted LFM radar signal are not damaged, and the two-dimensional modulation mode has a higher information transmission rate compared with a conventional one-dimensional modulation mode under the same bandwidth.

Description

Technical Field

[0001] This invention relates to the field of radar-communication integration, and more particularly to a radar-communication integrated system. Background Technology

[0002] With the rise of 5G / 6G, autonomous driving, electronic warfare and other fields, the demand for spectrum sharing and hardware reuse is becoming increasingly urgent. Therefore, radar-communication integration has become a research hotspot. Radar-communication integration aims to achieve efficient spectrum utilization, reduce system complexity, and meet the needs of future intelligent sensing and communication integration.

[0003] Currently, research on the design of integrated radar-communication waveforms mainly falls into two directions. The first is designing integrated waveforms based on communication signals. This approach primarily uses traditional communication waveforms, then adjusts the waveform structure or adds additional signal processing modes to enhance sensing capabilities. The second is designing integrated waveforms based on radar signals, aiming to utilize radar waveforms to achieve secondary communication functions. Unlike communication waveforms that directly carry modulation information, classic radar waveforms typically do not contain information signals; instead, they extract potential information or parameters from the echoes reflected from different targets. Therefore, this design is usually achieved by embedding communication information into the radar waveform while ensuring the primary detection function.

[0004] Currently, integrated waveforms designed based on radar signals are mostly based on linear frequency modulated signals (LFM), and communication information is embedded in them using phase modulation or amplitude modulation. However, this method is difficult to achieve high communication data rates while maintaining the ideal characteristics of LFM signals (including constant envelope and phase continuity). Summary of the Invention

[0005] In view of this, it is necessary to provide an integrated radar and communication system to solve the problem that existing technologies cannot achieve high communication data rates while maintaining the ideal characteristics of LFM signals.

[0006] To address the above problems, the present invention provides an integrated radar and communication system, comprising: an integrated signal transmitter; the integrated signal transmitter is used for: Get N Bit information stream; Sure N The first bit of information stream N The target initial frequency mapped by one bit of information, and the determination N The last bit of information in the bit stream N The target frequency modulation (FM) is mapped by 2 bits of information; among which, N 1 and N The sum of 2 equals N, N 1. N 2 and N All are positive integers; The target LFM radar signal is generated and transmitted based on the target initial frequency and the target modulation frequency.

[0007] The beneficial effects of this invention are: This invention modulates a single LFM radar signal using an integrated signal transmitter. N The bit information stream is then transmitted to the receiving end. During the modulation process, a two-dimensional modulation method using initial frequency modulation and frequency modulation is adopted. This modulation method does not destroy the constant envelope and phase continuity of the LFM radar signal. Moreover, tests have shown that, under the same bandwidth, this two-dimensional modulation method has a higher information transmission rate than the conventional one-dimensional modulation method.

[0008] Furthermore, different modulation methods can affect various performance characteristics of signal processing systems, such as communication rate, radar detection performance, and bit error rate during demodulation. In this invention, the number of bits under two degrees of freedom (initial frequency and modulation frequency) is... N 1 and N 2) The allocation can be freely adjusted, and the modulation method is more flexible. Therefore, the modulation method of the present invention can be adapted to different application scenarios and has stronger applicability. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a schematic diagram of a structure of an embodiment of an integrated radar and communication system provided by the present invention; Figure 2 A block diagram of a two-dimensional modulation system for an integrated signal transmitter provided by the present invention; Figure 3 (a)-(b) are the bit error rate-signal-noise ratio curves under different frequency modulation intervals and the bit error rate-signal-noise ratio curves under different initial frequency intervals provided by the present invention, respectively. Figure 4 (a)-(c) are the convergence curves of the ABA-CA algorithm PLSR, LSNR, and IPAE parameters under different Gaussian kernels provided by the present invention; (b)-(d) are the convergence curves of the ABA-CA algorithm PLSR, LSNR, and IPAE parameters under different cutoff thresholds provided by the present invention. Figure 5(a)-(c) are the time-domain / frequency-domain representations of the pulse filter outputs and the spectral images of the ideal filter output template when the ABA-CA algorithm iteration number n=0 provided by the present invention, respectively; Figure 6 (a)-(c) are the time-domain / frequency-domain representations of the pulse filter outputs and the spectral images of the ideal filter output template when the ABA-CA algorithm provided by this invention has an iteration number n=3, respectively; Figure 7 (a)-(c) are the time-domain / frequency-domain representations of the pulse filter outputs and the spectral images of the ideal filter output template when the ABA-CA algorithm provided by this invention has an iteration number n=6, respectively; Figure 8 The pulse compression result of the signal using the optimization method and ABA-CA algorithm provided by this invention is shown in the figure. Figure 9 (a)-(b) are comparison diagrams of pulse compression results of matched filtering and ABA-CA mismatched filter provided by the present invention, respectively; Figure 10 (a)-(b) are the range Doppler images of matched filtering and ABA-CA mismatch filtering under clutter conditions provided by the present invention, respectively; Figure 11 (a)-(b) are the three-dimensional and two-dimensional range Doppler images of the matched filter provided by the present invention, respectively; (c)-(d) are the three-dimensional and two-dimensional range Doppler images of the ABA-CA mismatch filter provided by the present invention, respectively. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0012] In the description of embodiments of the present invention, unless otherwise stated, "a plurality of" means two or more. The reference to "embodiment" herein means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0013] Before demonstrating the embodiments, the following terms will be explained.

[0014] LFM signal: A signal whose frequency changes linearly with time, widely used in radar, sonar, communications and other fields.

[0015] Constant envelope: refers to a signal whose amplitude (i.e., envelope) remains unchanged in the time domain and does not change with time.

[0016] Phase continuity: refers to the smooth change of the phase of a signal over time, without any jumps.

[0017] Initial frequency: The instantaneous frequency of the LFM signal at the start of the signal.

[0018] Frequency modulation: The rate at which frequency changes over time.

[0019] Bandwidth: The range of signal frequency variation, that is, the difference between the highest and lowest frequencies.

[0020] Duration: The duration of the signal.

[0021] Mismatched filters are a technique in radar signal processing that actively introduces a certain degree of "mismatch" (i.e., not perfectly matching the input signal) to reduce the sidelobe level of the matched filter's output signal while maintaining a high main lobe peak. Matched filters are the optimal filters for maximizing the output signal-to-noise ratio (SNR), but their output may have high sidelobes.

[0022] Out-of-band attenuation vector: used to describe the attenuation characteristics of a signal outside the target frequency band. Each element in the vector corresponds to the attenuation value (unit: dB) at a different frequency point.

[0023] Gaussian kernel: also known as radial basis function kernel, is a similarity measure based on Gaussian function.

[0024] This invention provides an integrated radar and communication system, which will be described in detail below.

[0025] Reference Figure 1 The diagram illustrates a structural schematic of an embodiment of the radar-communication integrated system provided by the present invention. System 10 includes: an integrated signal transmitter 110; the integrated signal transmitter 110 can be used for: Get N Bit information stream.

[0026] Sure N The first bit of information stream N The target initial frequency mapped by one bit of information, and the determination N The last bit of information in the bit stream N The target frequency modulation (FM) is mapped by 2 bits of information; among which, N 1 and N The sum of 2 equals N, N 1. N 2 and N All are positive integers.

[0027] The target LFM radar signal is generated based on the target's initial frequency and target modulation frequency.

[0028] This embodiment describes the LFM radar signal modulation process. Through this embodiment, the information flow can be modulated according to... N Bits are grouped and each group of information streams is modulated as a symbol onto each pulse of the LFM radar signal to achieve the generation of integrated radar communication signals.

[0029] In this embodiment, the number of all waveforms in the waveform set of the LFM radar signal is: The first one in the waveform set. The expression for each waveform is:

[0030] In the formula, It is a rectangular pulse function. It is a complex exponential function. Indicates the first k The initial frequency of the waveform, Indicates the first k The modulation frequency of the waveform, t Indicates time, T This represents the pulse width, where j is the imaginary unit.

[0031] In the formula, it is assumed that the preceding part of a single symbol is... The bit information stream maps the LFM radar signal's first bit. An initial frequency, then The bit information stream maps the LFM radar signal's first bit. Individual frequency, then and satisfy:

[0032] And the first The initial frequency and modulation frequency of the waveform are given by the following formula:

[0033] In the formula, and These are the intervals between the initial frequency and the frequency modulation, respectively.

[0034] For the Each symbol has the following structure:

[0035] In the formula, and They represent the first The initial frequency component and the first Each frequency modulation component is a bit of information corresponding to the initial frequency component or the frequency modulation.

[0036] Observing the spectral characteristics of the LFM radar signal in this embodiment from the frequency domain perspective, it can be seen that each waveform in the waveform set maintains a unique spectral structure, which is manifested in that each waveform has a different initial frequency and occupies a different bandwidth. For the frequency modulation scheme, since the pulse width of each pulse is kept constant, frequency modulation is equivalent to bandwidth modulation. That is:

[0037] In the formula, For bandwidth.

[0038] Let the initial frequency set and the frequency modulation set be:

[0039] The total bandwidth occupied by the waveform set consists of two parts: the signal spectrum spread caused by the initial frequency modulation. and the maximum bandwidth generated by frequency modulation Therefore, the total bandwidth occupied by the waveform set is:

[0040] The frequency band range is:

[0041] Clearly, modulation of both the initial frequency and the frequency modulation (FM) consumes additional bandwidth, resulting in lower bandwidth utilization in this embodiment compared to modulation schemes that do not require additional bandwidth for the initial phase, etc. However, this embodiment offers a more flexible modulation method within the available bandwidth. Furthermore, a fixed time width can be used. Thus, the appropriate frequency modulation can be selected. This allows for effective control over bandwidth expansion. Furthermore, the initial frequency set and the frequency modulation set jointly determine the spectral structure of the waveform set, which will affect the performance of the co-designed mismatched filter bank.

[0042] In summary, the LFM radar signal modulation method of this embodiment does not destroy the constant envelope and phase continuity of the LFM radar signal. Moreover, tests have shown that, under the same bandwidth, this two-dimensional modulation method has a higher information transmission rate than the conventional one-dimensional modulation method, and the modulation method is more flexible.

[0043] Reference Figure 2This diagram illustrates a two-dimensional modulation system block diagram of an integrated signal transmitter provided by the present invention. The integrated signal transmitter includes: a serial-to-parallel (S / P) converter, a bit packetization module, an initial frequency set mapping module, a frequency modulation set mapping module, and an LFM waveform generator. First, the serial data stream is converted into a two-dimensional modulation system by the S / P converter. N Bit-parallel sequence. Then the bit-blocking module will... N The bit parallel sequence is divided into N 1st group and N Two digits are grouped together. Then, the initial frequency set mapping module and the frequency modulation set mapping module are used to map them to the corresponding values ​​in the initial frequency set and the frequency modulation set, respectively, to obtain... N 1st group and N The target initial frequency corresponding to the 2-bit group and target frequency Finally, the LFM waveform generator is based on the mapping parameters. and The target LFM radar signal corresponding to the target's initial frequency and target modulation frequency is synthesized.

[0044] In some embodiments of the present invention, the radar-communication integrated system 10 further includes a communication receiver 120; the communication receiver 120 can be used for: The target LFM radar signal is received and demodulated by fractional Fourier transform to obtain the target initial frequency and target modulation frequency of the target LFM signal.

[0045] Determine the mapping between the target initial frequency and the target modulation frequency. N Bit information stream.

[0046] This embodiment describes the demodulation process of the target LFM radar signal. Since the LFM radar signal has excellent energy concentration characteristics in the fractional Fourier domain, the fractional Fourier transform (FRFT) can be used to estimate the parameters (initial frequency and modulation frequency) of the LFM radar signal. Based on the estimated parameters, the corresponding parameters of the LFM radar signal are obtained. N Bit information stream.

[0047] Specifically, for an arbitrary signal Its FRFT integral form is:

[0048] in The kernel function for FRFT is expressed as follows:

[0049] In the formula, Let be the transform order of the FRFT, which is related to the rotation angle. The relationship is:

[0050] The FRFT can be viewed as a counterclockwise rotation of the signal coordinates around the origin in the time-frequency plane. The standard Fourier transform can be seen as a special form of the fractional Fourier transform. When At this point, the FRFT degenerates into the standard Fourier transform. The axis is the frequency axis.

[0051] For LFM radar signals, at a specific rotation angle The transformation result exhibits a spike, similar to the frequency domain characteristics of a single-frequency signal, which resemble an impulse function. Since the optimal transformation of LFM radar signals with different modulation frequencies corresponds to a specific rotation angle, and different initial frequencies of LFM radar correspond to the horizontal axis coordinates of the fractional-order domain spikes appearing at that rotation angle, this makes it possible to demodulate the signal using the FFRFT method. Let the initial frequency of the LFM radar signal be... The frequency of modulation is Then the optimal FRFT transform order and peak coordinates are:

[0052] Therefore, by determining the optimal FRFT transform order and the peak coordinates under that transform, the initial frequency and modulation frequency of the LFM radar signal can be estimated.

[0053] In some embodiments of the present invention, the communication receiver 120 may specifically be used for: Obtain the frequency modulation set; the frequency modulation set includes different N The modulation frequency mapped by 2 bits of information.

[0054] The range of values ​​for the fractional Fourier transform order is determined by the maximum and minimum values ​​of the frequency modulation frequencies in the frequency modulation set.

[0055] The peak energy of the target LFM radar signal in the fractional Fourier domain is detected after performing a fractional Fourier transform on the target LFM radar signal according to each transform order within the range of values.

[0056] Based on the location of the highest energy peak and the transformation order, determine the target initial frequency and target modulation frequency of the target LFM signal.

[0057] There is a direct mathematical relationship between the optimal transform order and the frequency modulation frequency. Therefore, based on the mathematical relationship between the two and the range of values ​​for the frequency modulation frequency, the range of values ​​for the optimal transform order can be calculated.

[0058] Detection of each transformation order within the range of values The energy peak on the plane is recorded, and the optimal transformation order at the highest peak is also recorded. and their corresponding peak positions .

[0059] pass The estimated values ​​of the target's initial frequency and target modulation frequency are calculated using the following formulas:

[0060] In the formula,

[0061] Finally, the difference between the estimated value and each element in the set initial frequency set and frequency modulation set is calculated, and the element with the smallest difference is found as the final determined target initial frequency and target frequency modulation.

[0062] In this embodiment, since the search for the energy peak location and its order of occurrence is performed simultaneously, the demodulation scheme of this two-dimensional modulation scheme does not require increased system complexity compared to modulating the frequency modulation alone. This increases the deployability of this embodiment in practical engineering applications. It should be noted that in actual simulations, when using the above formula to estimate the initial frequency and frequency modulation of the LFM radar signal, dimensional normalization is required to restore the true parameter values.

[0063] In some embodiments of the present invention, the radar-communication integrated system 10 further includes: a radar receiver 130; the radar receiver 130 can be used for: Receive target LFM radar signals; Obtain the waveform set of the LFM radar signal, and determine the filtered output of each waveform based on the mismatch filter corresponding to each waveform in the waveform set.

[0064] The objective function is constructed based on the deviation between the filtered output and the ideal filtered output.

[0065] The objective function is solved to obtain the target design parameters for each mismatched filter, and the parameters of the mismatched filter are adjusted according to the target design parameters.

[0066] After the parameters are adjusted, the target LFM radar signal is filtered according to the target mismatch filter corresponding to the target LFM radar signal, and radar detection information is extracted from the filtered target LFM radar signal. The radar detection information may include the target's range information, speed information, etc.

[0067] In this embodiment, the number of waveforms in the waveform set is ,in , This represents the number of bits of the information stream modulated on a single LFM radar pulse signal. Let the number of bits in the waveform set be... The time-domain representation of a waveform signal is as follows: To ensure consistent output across pulse filtering, a corresponding mismatch filter needs to be designed for each waveform signal, thus requiring a set of mismatch filter banks to be designed based on the waveform set. Let the first... i The kernel function of the mismatch filter corresponding to each signal is: In the actual simulation, digital signal processing is used to sample and process the signal. Let the vector length of the signal be... Generally, the vector length of the mismatch filter kernel function is an integer multiple of the signal vector length; here, we let it be equal to the signal vector length. .

[0068] Let the first The signals and their corresponding mismatch filter kernel function vectors are as follows:

[0069]

[0070] No. i The filtered output of each pulse is:

[0071] in For signal The corresponding circular convolution matrix is ​​expressed as follows:

[0072] The optimization problem is modeled as a least-squares problem, aiming to find a set of mismatched filters such that the filtered outputs of different signals are as close as possible to the same ideal template. Let the template of the ideal filtered output be... , The objective function can be set by human experience or derived through an algorithm, then:

[0073] In the formula, 2 represents the 2-norm operation.

[0074] In an ideal situation, It should be a template that satisfies low sidelobes and narrow main lobes. Since the objective function is to calculate the minimum L2 error between the filtered output of all signals and the same template, the established objective function implicitly optimizes the performance parameters of the mismatched filter bank. That is, by solving the objective function, the optimal design parameters (target design parameters) of each mismatched filter can be obtained.

[0075] In summary, this embodiment reduces the impact of information flow loading on LFM radar signal detection performance by designing a mismatch filter bank corresponding to the waveform set.

[0076] Interference from communication information on radar detection stems from the fact that traditional radar signal processing methods treat deterministic signals. The modulation of communication information introduces randomness into the radar waveform, thus affecting the performance of radar signal processing. Traditional radar signal processing uses matched filtering to compress the echo signal pulses. Different waveforms between pulses produce different pulse compression responses, affecting the coherence between pulses. This effect is called Range Sidelobe Modulation (RSM), meaning that the sidelobe structure of the radar's pulse compression output is affected by information modulation, thus degrading radar detection performance. In radar detection, the RSM effect manifests as the range sidelobes of strong targets fluctuating randomly with changes in communication symbols. When multiple targets are present, the main lobe of a weak target is easily overwhelmed by the random sidelobes of strong targets, resulting in a significant decrease in detection performance.

[0077] Existing studies have largely failed to adequately assess the impact of the RSM effect, resulting in unpredictable fluctuations in the detection performance of designed waveform sets in practical applications.

[0078] Therefore, in some embodiments of the present invention, the radar receiver 130 may specifically be used for: Based on the deviations between the various filter outputs, a first constraint condition is constructed to constrain the consistency of each filter output.

[0079] Solve the objective function based on the first constraint condition.

[0080] In this embodiment, an ideal mismatch filter bank without RSM effect should satisfy the requirement that the filtered output of each pulse is completely consistent, that is:

[0081] However, the above equation cannot be fully satisfied unless a mismatched filter is used. h It is a wireless impulse response (IIR) filter. Therefore, this difference can be measured by constructing an inter-pulse filter output average error function:

[0082] The first constraint added for the RSM effect can then be:

[0083] In the formula, This indicates the maximum acceptable average error value between pulses.

[0084] This embodiment solves the objective function based on the first constraint condition, and designs a mismatch filter based on the solution result. This can ensure that the filtered outputs of each signal in the waveform set are as consistent as possible after being filtered by the mismatch filter, thereby eliminating the uncertainty brought about by information flow loading.

[0085] Solving the objective function implicitly involves optimizing various performance parameters of the mismatched filter bank. However, the optimization of these parameters comes at the cost of signal-to-noise ratio gain loss. Therefore, it is necessary to add constraints to some performance indicators to control this loss.

[0086] Therefore, in some embodiments of the present invention, the radar receiver 130 may specifically be used for: Determine the average signal-to-noise ratio gain loss for each filter output.

[0087] Based on the average signal-to-noise ratio gain loss, a second constraint condition is constructed to constrain the signal-to-noise ratio gain loss of each filter output.

[0088] Solve the objective function based on the second constraint.

[0089] In this embodiment, let the frequency domain representations of the signal and the mismatch filter be respectively... , For an arbitrary mismatch filter, the signal-to-noise ratio (SNR) of the signal output through that mismatch filter is:

[0090] In the formula, The one-sided power spectral density of the noise. Let be the vector length of the signal. For matched filtering, we have... , Let S(k) be the conjugate, then:

[0091] Therefore, the expression for the signal-to-noise ratio gain loss is:

[0092] By Cauchy-Schwarz inequality:

[0093] If and only if When the equality holds, the match sign is true. Therefore, we can conclude that matched filtering has the maximum output signal-to-noise ratio. It is a value greater than or equal to 1. Using any filter other than a matched filter is a mismatch filter, which will produce an appropriate loss of signal-to-noise ratio gain.

[0094] The expression for signal-to-noise ratio gain loss is transformed into the time domain. For a known signal, its energy... It is certain. According to Paswell's theorem, the energy of a signal in the time domain is equal to its energy in the frequency domain.

[0095] The time-domain expression for the signal-to-noise ratio gain loss is:

[0096] For the design of mismatched filter sets, it is necessary to consider The average signal-to-noise ratio gain loss of each waveform. In this invention, the signal used is of constant envelope, and the pulse width of each signal is... Certainly. Therefore, for all signals, their energy is equal, which is... .but The average signal-to-noise ratio gain loss for each pulse is:

[0097] For the first The peak output power of each waveform passing through the corresponding mismatch filter. Let... To determine the maximum acceptable average signal-to-noise ratio (SNR) loss, the constraint on the SNR gain loss is constructed as follows:

[0098] Due to peak output power It is an optimization variable For the relevant quantities, consider adding additional constraints to simplify the form. ,have:

[0099] Will The value of is normalized to a constant; let's take . At this point, the signal-to-noise ratio gain loss function is only related to... Related. Add the following constraints:

[0100] structure dimensional vector The vector is in The maximum value of the vector is 1 at its index, and all other elements are 0. Its function is to extract the peak power point of the filtered signal output. This constraint can be rewritten as:

[0101] In summary, the second constraint regarding the signal-to-noise ratio gain loss is:

[0102] This embodiment solves the objective function based on the second constraint condition, and designs a mismatch filter based on the solution result. This ensures that the signal-to-noise ratio gain loss of each signal in the waveform set is kept within a controllable range after the mismatch filter is processed.

[0103] In some embodiments of the present invention, the radar receiver 130 can be used to solve an objective function based on a first constraint and a second constraint. Specifically, it can solve the following model:

[0104] The problem described above is a convex optimization problem. Furthermore, due to the quadratic nature of the objective function and some constraints, this problem can be classified as a quadratic constrained quadratic programming (QCQP) problem, which can be solved using the CVX toolbox. The CVX toolbox is a MATLAB-based convex optimization modeling system for solving convex optimization problems, supporting various types such as linear programming, quadratic programming, and semidefinite programming.

[0105] The computational efficiency of traditional convex optimization methods is greatly limited by the number of signal sampling points. and the number of signals in the waveform set M The impact. In communication scenarios. and M The number of these numbers is usually quite large, and the time cost of such calculations is often immeasurable.

[0106] Therefore, in some embodiments of the present invention, the radar receiver 130 can also be used for: In each iteration, the average value of the filtered output is used as the ideal filtered output. Then, the ideal filtered output is updated by performing element-wise product calculation on the out-of-band decay vector and convolution calculation on the ideal filtered output using a Gaussian kernel.

[0107] Determine the deviation between the filtered output and the ideal filtered output after the update, and adjust the parameters of the mismatched filter based on the deviation.

[0108] When the iteration termination condition is met, the target design parameters of each mismatched filter are output.

[0109] Considering that the modulation mode used in this invention results in a special structure in the signal spectrum, it is more appropriate to design an algorithm for solving the objective function in the frequency domain. Therefore, this embodiment proposes an Adaptive Bandwidth Adjustment Cyclic Algorithm (ABA-CA). The specific implementation process of the ABA-CA algorithm is as follows: Let the waveform set be the first The frequency domain expression of each signal and its corresponding mismatched filter is as follows: and Then we have:

[0110] in for A 3D Discrete Fourier Transform (DFT) matrix. For the filtered output, we have:

[0111] In the formula, Represents the element-wise product of vectors.

[0112] To minimize the differences in the filtered outputs of various waveforms, some researchers have proposed a Joint Least Squares Mismatch Filter (JLS-MMF) based on the Least Squares (LS) algorithm, using the mean of the filtered outputs of each waveform as a template for the next iteration. This effectively reduces the RSM effect. The algorithm generated in this invention... It can be:

[0113] In order to adapt to the modulation mode proposed in this invention, The choice of frequency should be adjusted based on the specific frequency band characteristics of the waveform set. As mentioned above, the spectral structure of the waveform set is represented by multiple LFM spectra with different initial frequencies and bandwidths. Accordingly, the ABA-CA algorithm introduces an out-of-band attenuation vector. . The expression is as follows:

[0114] In the formula, k Represents the first waveform in the waveform set. k A waveform, This represents the common portion of the waveform set's spectrum. αThe average main lobe widening factor is the average bandwidth of each waveform signal in the waveform set after compression. The ratio of .

[0115] Restricted The frequency band range is the intersection of the spectra of all signals in the waveform set. Considering the consistency of the filtered output, all signals in the waveform set must produce an approximate frequency band structure after filtering. The introduction of this feature can maximize the preservation of signal energy and minimize the main lobe broadening after signal filtering.

[0116] With out-of-band attenuation vector Related The update strategy is as follows:

[0117] For traditional matched filtering methods, the peak-to-sidelobe ratio (PSLR) after pulse compression of the LFM signal is approximately -13 dB. This high sidelobe characteristic makes matched filtering easily overwhelmed by the sidelobes of strong targets when detecting weak targets. Consider... The same considerations must be taken into account when dealing with the issue of spectrum shaping. Its low sidelobe characteristics. Low sidelobe... In the frequency domain, it manifests as It possesses a smooth spectral envelope, which can be explained by the spectral characteristics of rectangular pulse signals and Gaussian pulse signals. The time-domain expression of a Gaussian pulse is:

[0118] Its corresponding frequency domain expression is:

[0119] Gaussian pulses exhibit low sidelobe characteristics in the time-frequency domain. Therefore, for In terms of spectrum shaping, it is necessary to suppress the rectangular wave components in the spectrum as much as possible, making its envelope close to a Gaussian function, thereby achieving... The low sidelobe characteristics. One feasible method is to use a Gaussian kernel G vector pair. Perform convolution operations:

[0120] In the formula,

[0121] In the formula, Standard deviation, The length of the Gaussian kernel, k The first Gaussian kernelk One sampling point.

[0122] It can be controlled and The value is used to balance the degree of sidelobe suppression with the signal-to-noise ratio gain loss.

[0123] After updating the ideal filter output using the out-of-band attenuation vector and Gaussian kernel, the deviation between the filtered output and the updated ideal filter output can be calculated, and the parameters of the mismatched filter can be adjusted based on the deviation.

[0124] Specifically, the deviation can be a correction vector. The optimization variables are corrected using this vector. The value of , thus making the filtered output approximate . Correction vector The expression is:

[0125] The process of adjusting the parameters of the mismatch filter is as follows:

[0126] After the mismatch filter parameters are adjusted, it can be determined whether the iteration termination condition is met. The iteration termination condition can be either the first constraint condition or the second constraint condition. If the iteration termination condition is met, the target design parameters of each mismatch filter are output; if the iteration termination condition is not met, the next round of iteration is performed.

[0127] In summary, this embodiment adaptively updates the suitable ideal filter output in each iteration. This makes No longer subject to prior human designation, thus greatly reducing the risk of... The algorithm overcomes the uncertainty introduced by the subjectivity of the selection process. Furthermore, while eliminating the RSM effect, it minimizes the broadening of the pulse compression main lobe structure by mismatch filtering, and exhibits good performance in terms of signal-to-noise ratio gain loss and peak-to-side-lobe ratio. Moreover, due to its low computational complexity, the design time cost of the mismatch filter bank is controllable. Therefore, this algorithm is particularly suitable for radar and communication integrated systems with large waveform sets.

[0128] In some embodiments of the present invention, in order to prevent the unbounded growth of the correction value, the radar receiver 130 can also be used for: If the deviation is less than or equal to the preset cutoff threshold, the parameters of the mismatch filter are adjusted based on the deviation. If the deviation is greater than the preset cutoff threshold, the parameters of the mismatch filter will be adjusted according to the preset cutoff threshold.

[0129] Specifically, the correction vector after applying the truncation threshold can be represented by the following formula:

[0130] In the formula, This is the truncation threshold.

[0131] Referring to Table 1 below, pseudocode for an ABA-CA algorithm implementation provided by this invention is shown. In this pseudocode, the initial value of LSNR is set to the minimum LSNR, and LSNR continuously increases during the iteration process. Therefore, when When that happens, exit the loop.

[0132] Table 1. Pseudocode for the ABA-CA algorithm implementation

[0133] As shown in Table 1, the algorithm of the present invention is simple to implement and has low computational complexity.

[0134] For the radar-communication integrated system of the present invention, the following sets of simulation experiments were designed to demonstrate the superiority of the method: (1) Communication performance analysis. Analysis method: The communication performance under the proposed modulation framework was simulated and analyzed. The good noise immunity of the proposed scheme was verified by Monte Carlo experiment. Analysis conclusion: Through comparative analysis, it was verified that the two-dimensional modulation scheme has a higher information transmission rate than the one-dimensional one-dimensional one-dimensional one-dimensional modulation scheme under the same bandwidth.

[0135] (2) Algorithm Performance Analysis. Analysis Method: The stable convergence characteristics of the algorithm under different parameter conditions were analyzed. Analysis Conclusion: By comparing the complexity and optimization effect with traditional convex optimization algorithms, it is proved that the algorithm proposed in this invention can greatly reduce the computational complexity while ensuring minimal performance loss. In addition, comparison with other mismatched filters demonstrates the excellent radar detection performance of the designed mismatched filter.

[0136] (3) Simulation analysis of radar echo signal processing. Analysis method: The simulation experiment simulated the reception of radar echo signals from multiple targets, and the corresponding signal processing was compared with that of a traditional matched filter. Analysis conclusion: The mismatch filter proposed in this invention can greatly reduce the radar detection performance degradation caused by communication information modulation.

[0137] The shared parameters used in the simulation experiment are shown in Table 2: Table 2 Simulation Parameter Table

[0138] The simulation experiments also used the following performance evaluation parameters: ① To quantitatively measure the degree of relief of the RSM effect, a parameter called interpulse mean error (IPAE) is defined. Its definition is:

[0139] in The mean of the filtered outputs of each pulse. This dimensionless metric reflects the similarity of the filtered outputs of the waveform set.

[0140] ② Define the normalized inter-pulse average error as:

[0141] in This represents the inter-pulse average error of the matched filter, used as a standard value for normalization.

[0142] ③ To effectively mitigate the RSM effect, ABA-CA truncates the common portion of the signal bandwidth. This results in the actual bandwidth used for radar resolution being smaller than the signal bandwidth, leading to main lobe broadening after pulse compression. The average main lobe broadening coefficient is given by the following formula:

[0143] In the formula, For the waveform set The bandwidth of a compressed signal pulse. When matched filtering is used. This indicates that matched filtering does not introduce additional widening to the pulse compression main lobe. In ABA-CA, .

[0144] The specific processes of the above three analyses are as follows: (1) Communication performance analysis The simulation experiment was conducted in an additive white Gaussian noise channel, using a single LFM pulse to modulate 4 bits, with 2 bits modulated at the initial frequency and the modulation frequency respectively.

[0145] Figure 3 (a) plotted in Under different conditions The bit error rate-signal-noise ratio (BER-SNR) curve. With... As the value increases, the BER-SNR curve shifts to the left, indicating that a larger value... This improves the distinguishability between adjacent signals. Greater distinguishability increases the noise immunity tolerance of the demodulation scheme, thus resulting in a lower bit error rate to some extent. Figure 6 (b) The drawing was made in Under different conditions The BER-SNR curve, and Figure 6 (a) They exhibited similar characteristics.

[0146] Overall, Figure 6 The curves demonstrate the excellent noise immunity of the FRFT-based demodulation scheme. Even in low signal-to-noise ratio (SNR) scenarios (SNR < -5dB), the integrated signal transmission maintains a low bit error rate. This good noise immunity is attributed to the energy concentration characteristics of the LFM signal in the fractional Fourier domain of a specific order.

[0147] In 3(b), with compared to, It improves the signal-to-noise ratio tolerance by approximately 2 dB under the same bit error rate. Under conditions of limited bandwidth resources, a smaller... and This means higher communication transmission rates. Therefore, the design of waveform sets can balance communication rate and noise immunity performance according to actual needs.

[0148] Compared to one-dimensional modulation schemes using a single initial frequency or frequency modulation, the two-dimensional modulation framework proposed in this invention significantly increases the transmission rate of communication within the same frequency band. For simplified analysis, it is assumed that the number of bits at the initial frequency and the frequency modulation is equal, i.e. The bandwidth occupied by the initial frequency interval is equal to that occupied by the frequency modulation interval, that is: The analysis of the waveform set bandwidth above shows that the total bandwidth occupied under this assumption is... for:

[0149] Under this bandwidth constraint, the number of bits transmitted per pulse in the two-dimensional modulation framework is For one-dimensional modulation, it is With large waveform sets For example, when the radar pulse repetition period is... At that time, the communication transmission rate can reach This method shows promising application prospects in military radar communication integration scenarios with high detection requirements and relatively low communication rate demands. In comparison, the communication transmission rate of a one-dimensional modulation scheme with the same bandwidth is 60 kb / s. The proposed solution achieves a 66.7% rate increase. This approach is applicable to waveform sets... The rate increase is more pronounced when the value is larger.

[0150] (2) Algorithm performance analysis The low complexity of the ABA-CA algorithm stems from its frequency domain processing architecture and fast convergence. Because it can be accelerated using FFT, the algorithm's complexity is proportional to the size of the waveform set. It exhibits a linear relationship with the signal dimension. It exhibits an approximately linear relationship, and its overall complexity is... Considering its empirically observed convergence within 4-8 iterations, this high design efficiency enables real-time processing of large-scale waveform sets, while traditional optimization methods are computationally extremely costly. As shown in Table 3, compared with the QCQP solver (internal point method), ABA-CA performs better in typical radar communication scenarios (…). , This reduces the number of operations by nine orders of magnitude.

[0151] Table 3 Comparison of Algorithm Complexity

[0152] Figure 4 The following are examples of different Gaussian kernels G and cutoff thresholds. Convergence characteristics of various mismatch filter performance indicators in the ABA-CA MMF. Figure 7 (a)-(c) demonstrate in Select different conditions The convergence characteristics of each parameter of the ABA-CA algorithm were observed during the experiment. Simulation results show that... The main factor affecting the convergence speed of the algorithm is selecting a larger [size / size]. This ensures that ABA-CA converges stably with fewer iterations. The iterative curves show that ABA-CA maintains stable convergence under different parameter conditions, and effectively suppresses sidelobe levels and maintains a small signal-to-noise ratio loss while greatly improving the consistency of the inter-pulse filter output.

[0153] Figure 4 (d)-(f) show the cutoff threshold The convergence characteristics of the ABA-CA algorithm were investigated by selecting four Gaussian kernel vectors with different lengths and standard deviations under the given conditions. Simulation results show that the ABA-CA algorithm can converge stably under different G conditions, and the selection of G mainly affects the PSLR and LSNR of the mismatched filter. By reasonably selecting the Gaussian kernel vector G, the ABA-CA algorithm can achieve an effective trade-off between PSLR and LSNR.

[0154] Figure 5-7 They were shown respectively n =0, n =3, n The time-domain / frequency-domain representations of the pulse filter outputs of the ABA-CA algorithm at time = 6 and the spectrum image of the ideal filter output template are shown below. n The iteration count is defined as follows: each pulse comprises four waveforms selected from a set of 16 waveforms, each with a different initial frequency and modulation frequency. As the algorithm iterates, the filtered outputs of the different waveforms gradually converge, which is consistent with... Figure 4 The IPAE curve shown corresponds to this. Furthermore, the iterative process also reduces PSLR, in n At 6, it reached a level of approximately -45dB.

[0155] Table 4 Performance of mismatched filters using optimization methods and ABA-CA algorithm

[0156] Since ABA-CA is an approximate algorithm for solving convex problems, it incurs a performance loss compared to the optimization algorithm. Table 4 shows a performance comparison of mismatched filters designed using the optimization algorithm and ABA-CA, using the interior-point method commonly used in CVX for solving QCQP problems for comparison. To verify the extent of the performance loss of ABA-CA and ensure the validity of the comparison, the same constraint values ​​as the convergence values ​​of the ABA-CA algorithm were applied to inter-pulse error and signal-to-noise ratio loss. Furthermore, regarding the selection of the optimal template... Selected after convergence with ABA-CA The main lobe portion. The performance loss is primarily reflected in the degree of suppression of the side lobes. Figure 8 This performance difference is further demonstrated. Simulation experiments show that ABA-CA suffers less performance loss compared to optimization algorithms, with a loss of less than 10dB on PSLR.

[0157] Table 4 further compares the performance differences resulting from using different ideal templates in the traditional optimization method. Three templates with different main lobe widths were selected for the experiment. , , The main lobe width decreases sequentially, and the same inter-pulse error and LSNR constraint values ​​are added. Simulation results show that... The main lobe width has a significant impact on the optimization effect. An excessively narrow main lobe width causes the mismatched filter to generate gain outside the signal band, thus increasing the LSNR. Since the same LSNR constraint is applied here, the performance difference is reflected in the decrease in PSLR. Compared to using... ,use The PSLR value decreased by nearly 30 dB. This performance difference further demonstrates that ABA-CA can adaptively adjust... Its superiority greatly reduces subjective choice. Despite the uncertainty, it maintained good performance.

[0158] Table 5 Performance Comparison of Different Mismatched Filters

[0159] Table 5 shows the performance parameters of different pulse compression filters. Compared with matched filters and their windowed pulse compression modes, ANA-CA MMF significantly reduces IPAE, which can effectively alleviate the RSM effect caused by communication information modulation. In addition, ANA-CA MMF achieves a significant reduction in PSLR at a relatively small LSNR cost, comparable to the sidelobe level of traditional windowed matched filters.

[0160] Table 5 also compares ABA-CA MMF with other mismatch filter design methods. LS-MMF does not significantly reduce the RSM effect. JLS-MMF and JMMF-CA (Joint Mismatch Filter with Cyclic Adjustment) reduce IPAE to some extent and improve the RSM effect by updating the iterative template with the mean of each pulse filter output, but their time-domain design orientation does not fit well with the waveform set with the special spectral structure proposed in this paper. In contrast, the frequency-domain-oriented design of ABA-CA MMF is well adapted to the proposed modulation scheme, and its IPAE can be reduced to about -300dB. This difference in inter-pulse filter output is almost negligible in practical applications. Therefore, ABA-CA MMF makes a significant improvement in eliminating the RSM effect. In addition, compared with other mismatch filters, ABA-CA MMF maintains the maximum common bandwidth energy of the waveform set, achieves a smaller LSNR at the cost of a smaller increase in the main lobe broadening coefficient, and maintains high performance in sidelobe level and RSM effect improvement.

[0161] (3) Radar echo signal analysis The experiment simulated multiple moving targets at different distances and speeds in space, and generated corresponding echo signals based on the targets. These echo signals were then subjected to pulse compression and Doppler processing at the radar receiver. By comparing the matched filter and the mismatched filter bank designed using the proposed algorithm in different environments, the superior performance of the ABA-CA MMF in radar detection was verified.

[0162] Pulse compression processing was performed on the echoes of three nearby targets under Gaussian white noise environment with SNR=10dB. Figure 9(a) and (b) show the pulse compression processing results using MF and ABA-CA MMF, respectively. When the target distances from the radar are 5450m, 5500m, and 5600m, the main lobe of target 1 (weak target) is overwhelmed by the sidelobes of target 2 (strong target) when using matched filtering, making target 1 undetectable. When using ABA-CA MMF, the PSLR reaches approximately -45dB, and the main lobe peaks of all three targets are clearly detected, demonstrating the detection performance of ABA-CA MMF against weak targets when detecting multiple nearby targets.

[0163] The experiment simulated the detection capability of the ABA-CA MMF for multiple targets under strong clutter conditions and compared it with that of a traditional matched filter, just as... Figure 10 As shown, under conditions of SNR=10dB and SCR (signal-to-noise ratio)=-20dB, the simulation aimed to detect targets at three different distances and velocities. Figure 10 In (a), when using a matched filter, a large amount of clutter energy remains on the range-Doppler image (RDM). This is because the RSM effect causes inconsistent compression structures of multiple pulses, making it impossible to completely eliminate clutter energy during pulse cancellation. Target 3 (a weak target) is completely submerged in clutter and cannot be distinguished. Figure 10 (b) When pulse compression is performed using ABA-CA MMF, the IPAE of the filtered output can reach -300dB, which basically eliminates the RSM effect. Most of the clutter energy is canceled out, thus enabling all three targets to be clearly detected. Simulation results show that compared with matched filtering, ABA-CA MMF can improve the average suppression of clutter by nearly 60dB.

[0164] Figure 11 (a)-(d) demonstrate the improvement in Doppler coherence of ABA-CA MMF compared to MF under Gaussian white noise conditions. In this scenario, the experiment simulated the radar's ability to distinguish targets with the same range but different velocities. (Comparison) Figure 11 (b) and Figure 11 (d) At the range cell where target 3 is located, matched filtering produces significant banded Doppler blurring. This is because matched filtering cannot eliminate the additional phase effects introduced by communication information. In contrast, ABA-CA MMF greatly reduces inter-pulse averaging error and eliminates the effects of this additional phase. Figure 11 (c) The Doppler dimension of target 3 shows that it is a standard sinc function, which is basically the same as the effect when radar is processed without additional communication information.

[0165] The reduction in Doppler coherence, as one of the potential effects of the RSM effect, increases the sidelobes of the Doppler dimension, making it more difficult to detect targets with different velocities at the same range cell. Figure 11 In (b), the main lobe of target 2 is almost completely overwhelmed by the additionally enhanced sidelobes. Conversely, Figure 11 (d) The main lobe of target 2 is clearly visible under ABA-CA MMF, which increases the reliability of radar Doppler processing. Simulation results show that, under the set simulation environment, ABA-CA MMF can improve the average Doppler sidelobe suppression by about 6 dB compared to MF.

[0166] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0167] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A radar-communication integrated system, characterized in that, include: Integrated signal transmitter; The integrated signal transmitter is used for: Get N Bit information stream; Sure N The first bit of information stream N The target initial frequency mapped by one bit of information, and the determination N The last bit of information in the bit stream N The target frequency modulation (FM) is mapped by 2 bits of information; among which, N 1 and N The sum of 2 equals N , N 1. N 2 and N All are positive integers; The target LFM radar signal is generated and transmitted based on the target initial frequency and the target modulation frequency.

2. The radar-communication integrated system according to claim 1, characterized in that, It also includes a communication receiver; the communication receiver is used for: The target LFM radar signal is received, and the target LFM radar signal is demodulated by fractional Fourier transform to obtain the target initial frequency and the target modulation frequency of the target LFM signal; Determine the target initial frequency and the target modulation frequency mapped to the N Bit information stream.

3. The radar-communication integrated system according to claim 2, characterized in that, The communication receiving end is specifically used for: Obtain a frequency modulation set; the frequency modulation set includes different N The frequency modulation rate mapped by 2 bits of information; The range of values ​​for the fractional Fourier transform order is determined based on the maximum and minimum values ​​of the frequency modulation frequencies in the frequency modulation set. Based on the value range, the target LFM radar signal is demodulated by fractional Fourier transform to obtain the target initial frequency and the target modulation frequency of the target LFM signal.

4. The radar-communication integrated system according to claim 3, characterized in that, The communication receiving end is specifically used for: The energy peak value of the target LFM radar signal in the fractional Fourier domain is detected after performing a fractional Fourier transform on the target LFM radar signal according to each transform order within the range of values. Based on the location of the highest energy peak and the transformation order, the target initial frequency and the target modulation frequency of the target LFM signal are determined.

5. The radar-communication integrated system according to claim 1, characterized in that, Also includes: Radar receiver; the radar receiver is used for: Receive the target LFM radar signal; Obtain the waveform set of the LFM radar signal, and determine the filtered output of each waveform according to the mismatch filter corresponding to each waveform in the waveform set; Based on the deviation between the filtered output and the ideal filtered output, a target function is constructed; The objective function is solved to obtain the target design parameters for each of the mismatched filters, and the parameters of the mismatched filters are adjusted according to the target design parameters. After the parameters are adjusted, the target LFM radar signal is filtered according to the target mismatch filter corresponding to the target LFM radar signal.

6. The radar-communication integrated system according to claim 5, characterized in that, The radar receiver is specifically used for: Based on the deviations between the various filter outputs, a first constraint condition is constructed to constrain the consistency of the various filter outputs. The objective function is solved based on the first constraint.

7. The radar-communication integrated system according to claim 5, characterized in that, The radar receiver is specifically used for: Determine the average signal-to-noise ratio gain loss for each of the filter outputs; Based on the average signal-to-noise ratio gain loss, a second constraint condition is constructed to constrain the signal-to-noise ratio gain loss of each of the filtered outputs; The objective function is solved according to the second constraint.

8. The radar-communication integrated system according to claim 5, characterized in that, The radar receiver is specifically used for: In each iteration, the average value of the filtered output is used as the ideal filtered output. Then, the element-wise product of the ideal filtered output is calculated using the out-of-band decay vector, and the convolution calculation is performed on the ideal filtered output using a Gaussian kernel to update the ideal filtered output. Determine the deviation between the filtered output and the ideal filtered output after the update, and adjust the parameters of the mismatched filter according to the deviation; When the iteration termination condition is met, the target design parameters of each mismatched filter are output.

9. The radar-communication integrated system according to claim 8, characterized in that, The out-of-band attenuation vector expression is as follows: In the formula, k Represents the first waveform in the waveform set. k A waveform, This represents the common portion of the waveform set's spectrum. α The average main lobe widening factor is the average bandwidth of each waveform signal in the waveform set after compression. The ratio of .

10. The radar-communication integrated system according to claim 8, characterized in that, The radar receiver is specifically used for: If the deviation is less than or equal to a preset cutoff threshold, the parameters of the mismatch filter are adjusted based on the deviation. If the deviation is greater than the cutoff threshold, the parameters of the mismatch filter are adjusted using the cutoff threshold.

Images