Uplink ISAC system interference suppression method based on signal folding

Abstract

The invention provides an uplink ISAC system interference suppression method based on signal folding. The uplink ISAC system interference suppression method is used for dynamic range optimization and interference elimination in an uplink. According to the technology, a high-dynamic-range signal is compressed through a signal folding mechanism, folding operation is firstly carried out on a received signal, and the amplitude of the received signal is limited within a preset folding threshold interval; performing fast Fourier transform (FFT) operation on the target echo in the folding domain, and extracting a distance-speed (RV) spectrum of the target; obtaining an interference reference reconstruction signal according to the RV spectrum; and finally, estimating and eliminating the radar interference signal in the folding domain by adopting an adaptive filtering algorithm, and recovering a communication signal. According to the method, saturation distortion of the ADC is avoided, and reconstruction-free radar processing and adaptive interference suppression are realized in the folding domain.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of wireless communication and sensing fusion, and particularly relates to an interference suppression technology for integrated sensing and communication (ISAC). BACKGROUND

[0002] As one of the key technologies for 6G, ISAC systems are widely used in large-scale resource reuse scenarios, such as vehicle collision avoidance and pedestrian detection in intelligent transportation systems, real-time environment perception and path planning in autonomous driving, precision equipment monitoring and fault prediction in industrial Internet of Things, traffic flow optimization and public safety monitoring in smart cities, and non-contact vital sign monitoring in the field of medical health. These application scenarios often involve long-distance weak target sensing (such as low-altitude drones or remote pedestrians) and strong interference environments (such as building reflections or self-interference), requiring systems to have a dynamic range of more than 100 dB to support high-precision positioning (sub-meter level) and low-latency response (microsecond level), thereby improving overall spectrum efficiency and resource utilization.

[0003] In the integrated sensing and communication (ISAC) system, the interference problem of the uplink is a major challenge, especially when both sensing signals and communication signals are transmitted simultaneously. Traditional interference suppression methods usually use digital domain interference cancellation techniques, but these methods require the signal to be completely in the digital domain and assume that the dynamic range of the receiver is large enough. However, in actual systems, the dynamic range of the receiver is often limited, leading to signal truncation and distortion.

[0004] One existing interference suppression scheme is to cancel or significantly reduce interference through an interference cancellation system in a wireless network, for example, US patent USRE43791E1 (Systems and methods for interference cancellation) proposes an interference cancellation system and method for separating interference sources and suppressing continuous noise and random interference signals, thereby reliably filtering out useful echo signals from received signals. This scheme describes a method of gradually estimating and canceling interference through self-interference cancellation (SIC) technology, which is suitable for interference suppression in wireless communication systems, but it relies on a high dynamic range ADC to avoid signal saturation, and in the high dynamic scenario of ISAC uplink (such as a sensing signal dynamic range of 100 dB), it is easy to cause signal distortion due to insufficient ADC dynamic range (usually only 40 dB), which cannot effectively handle sensing self-interference or near-end reflection signals.

[0005] To extend the dynamic range of ADC, the prior art also adopts high dynamic range analog-to-digital conversion methods, for example, US patent US9654134B2 (High dynamic range analog-to-digital conversion) proposes a multi-stage analog-to-digital conversion system that uses window functions and shifting to match high-gain data frames with target data frames, achieving dynamic adjustment of dynamic range. This scheme describes in detail the processing of high dynamic signals through a multi-stage ADC architecture (such as the combination of coarse quantization and fine quantization) to avoid signal clipping distortion and is suitable for systems that require high-resolution conversion. However, this method requires precise analog circuit components, resulting in high hardware costs, increased power consumption, and difficulty in integration in large-scale MIMO array ISAC systems, which cannot meet the real-time interference suppression requirements in multi-user scenarios.

[0006] In addition, some solutions indirectly support interference suppression by enhancing the dynamic range of ADC, for example, US patent US11606100B2 (Method and apparatus for enhancing dynamic range in an analog-to-digital converter) proposes a device and method for improving the dynamic range of ADC through dynamic gain control and noise shaping. This scheme emphasizes the introduction of adaptive threshold adjustment during analog-to-digital conversion to process high dynamic input signals (such as radar echoes) and reduce quantization noise, but still requires subsequent digital processing for interference cancellation. In ISAC applications, this scheme can extend the dynamic range to a higher level (such as more than 100 dB), but has high computational complexity and strict requirements for hardware precision, which cannot avoid the overhead of signal reconstruction and is limited in performance in strong interference environments (such as self-interference power higher than the communication signal by several orders of magnitude).

[0007] In summary, existing digital domain interference cancellation techniques rely on the dynamic range of the receiver, and when the signal exceeds the dynamic range, it will cause signal truncation and distortion. Existing schemes for extending the dynamic range require complex signal reconstruction algorithms, increasing the computational complexity and hardware cost of the system. Currently, in the case of coexistence of strong sensing signals and weak communication signals, it is still difficult to effectively suppress interference, resulting in a decrease in the quality of communication signals. SUMMARY

[0008] The technical problem to be solved by the present application is to reduce the dynamic range requirement and computational complexity of the system, and to provide a radar signal interference cancellation scheme without signal reconstruction.

[0009] The technical scheme adopted by the present application to solve the above technical problems is an uplink ISAC system interference suppression method based on signal folding, comprising the steps of:

[0010] performing a folding operation on the received signal to limit its amplitude within a preset folding threshold interval;

[0011] performing a fast Fourier transform (FFT) operation on the target echo within the folding domain to extract a range-velocity (RV) spectrum of the target;

[0012] obtaining an interference reference reconstruction signal according to the RV spectrum; wherein, is an interference reference reconstruction signal at the kth moment, s rad is a radar transmission signal at the kth moment, m is a channel number, g m is an mth channel coefficient, Δf is a subcarrier spacing, l is a subcarrier index, μ' is a configuration parameter of the OFDM subcarrier spacing, T is a sampling period, is a time delay on the range dimension of the RV spectrum, is a Doppler shift on the velocity dimension of the RV spectrum;

[0013] performing estimation and elimination of the radar interference signal within the folding domain using an adaptive filtering algorithm to recover the communication signal.

[0014] The conventional method is difficult to effectively suppress interference in the case where a strong sensing signal and a weak communication signal coexist, resulting in a decline in the quality of the communication signal. The technical problem to be solved by the present application is to provide an interference suppression method without signal reconstruction, reducing the dynamic range requirement and computational complexity of the system. Radar signal processing and interference elimination are directly performed within the folding domain, avoiding the additional overhead brought by signal reconstruction. The interference suppression effect of the uplink ISAC system is improved, especially in the case where a strong sensing signal and a weak communication signal coexist.

[0015] Specifically, the folding operation is:

[0016]

[0017] wherein, M λ is a folding operation function, x is a signal input to the folding operation, λ is a folding threshold, represents a floor function.

[0018] Specifically, the adaptive filtering algorithm is used to estimate the radar interference signal within the folding domain, specifically:

[0019]

[0020] wherein, is an estimated radar interference signal at the kth moment, the input reference signal vector x[k] of the filter, H represents a conjugate transpose, and the interference reference reconstruction Composition, T denotes transposition, L is the length of the filter, and w is the adaptive filter weight.

[0021] Specifically, the adaptive filtering algorithm is used in the folding domain to eliminate the radar interference signal, and specifically is:

[0022]

[0023] Wherein, y λ [k]=M λ (y[k]) is the received signal that has been folded at the kth moment.

[0024] Specifically, the updating method of the adaptive filter weight is:

[0025]

[0026] w[k+1] represents the filter weight at the k+1th moment, and mu s is the set step size.

[0027] The beneficial effects of the present application are:

[0028] (1) Dynamic range optimization: through the folding operation, the signal dynamic range is limited in a fixed range, avoiding the signal truncation problem, and expanding the dynamic range of the system.

[0029] (2) Reduce system complexity: directly process the radar signal and eliminate the interference in the folding domain, without signal reconstruction, reduce the system complexity and hardware cost.

[0030] (3) Improve the interference suppression effect: through the folding domain LMS structure, effectively suppress the interference of strong sensing signal to weak communication signal, improve the interference suppression performance of the system. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 The present application is an interference suppression scheme for ISAC system;

[0032] Figure 2 The present application is an interference suppression scheme for ISAC system; DETAILED DESCRIPTION

[0033] The present application proposes an uplink ISAC system interference suppression method based on signal folding, as shown in Figure 1 .

[0034] The transmission (TX) and reception (RX) processing flow of the OFDM radar communication integrated system is divided into three parts: OFDM radar communication transmission, OFDM radar reception, and OFDM communication reception.

[0035] 1. OFDM radar communication transmission (OFDM radar communication TX)

[0036] From the bit stream Bit, the input bit stream is encoded, the encoded data is mapped to the quadrature amplitude modulation QAM symbol, the frequency domain QAM symbol is converted to the time domain signal through the inverse fast Fourier transform IFFT, the cyclic prefix CP is added to resist the multipath effect, and the digital signal is converted to the analog signal through the digital-to-analog conversion DAC. After being amplified by the power amplifier PA, the signal transmission is completed by the antenna. Part of the transmitted signal is used for detection (OFDM Radar), and the other part is used as uplink communication (OFDM Communication Uplink).

[0037] In the process of OFDM radar communication transmission and reception, radar interference RI caused by the transmission signal of the radar system to the receiving end of the communication system and self-interference SI directly leaked to the receiving end antenna through the unintended propagation path will occur.

[0038] 2. OFDM radar reception (OFDM radar RX)

[0039] After the receiving end antenna captures the signal, the signal is amplified through the low noise amplifier LNA, the received signal is processed in the folding domain, the analog signal after the folding domain processing is converted to the digital time domain signal, the cyclic prefix CP is removed, the time domain signal is converted to the frequency domain signal through the fast Fourier transform FFT, the frequency domain channel estimation or signal processing is performed through the least square LS, and the radial velocity of the radar target is processed in the folding domain to obtain the distance-velocity spectrum of the target.

[0040] At the same time, the time domain signal after removing the cyclic prefix CP flows to the OFDM communication RX, which is used for interference cancellation of the communication signal.

[0041] 3. OFDM communication reception (OFDM Communication RX)

[0042] The time domain signal after removing the cyclic prefix CP from the OFDM radar RX and the distance-velocity spectrum of the target are received to perform the folding domain self-interference SI / radar interference RI cancellation, that is, the signal is converted to the frequency signal through the FFT and the communication channel is estimated; then the frequency domain signal is reflected to the QAM symbol (DeQAM), and decoded to the bit stream Bit, completing the reception of the communication data.

[0043] 1. Folding domain processing: radar signal processing is directly performed in the folding domain, without signal reconstruction. Through the folding operation, the signal dynamic range is limited in a fixed range, avoiding the signal truncation problem. The mathematical expression of the folding operation is as follows:

[0044]

[0045] where M λ is the folding operation function, x is the signal inputted to the folding operation, λ is the folding threshold, denotes the floor function.

[0046] The range-velocity (RV) spectrum of the target is extracted directly in the folding domain by using the equivalence in the folding domain, without recovering the signal to the original domain.

[0047] Suppose the radar transmitted signal is s rad , after the radar channel gain g, the original received signal y is:

[0048]

[0049] where P rad is the transmitted power of the radar signal, n is the noise, is the target echo.

[0050] In the folding domain, the folding domain received signal y folded is expressed as:

[0051]

[0052] According to the properties of the folding operation, the folding domain received signal is decomposed as:

[0053]

[0054] where is the error term introduced by the folding operation, called the folding error, denotes the integer set, M denotes the total number of pulses transmitted by the radar in one coherent processing interval, and M×1 This means that e is a column vector, which contains the error introduced by the folding operation on each received pulse in one complete coherent processing interval. From the above equation, it can be seen that although the signal y is folded, the folded signal y folded can still be expressed as the original unambiguous target echo plus an error term e.

[0055] The fast Fourier transform (FFT) operation is performed on the folded signal y folded to obtain:

[0056]

[0057] Since e is a periodic function, its FFT result is mainly concentrated in the low frequency part with small amplitude. Therefore, FFT(e) has limited impact on the target's RV spectrum extraction. The target's RV spectrum is obtained by performing two FFT operations on the radar signal. The first FFT is used to extract the range information, and the second FFT is used to extract the velocity information. Due to the linear property of FFT, the FFT result in the folded domain is consistent with the FFT result in the original domain in phase information. Therefore, the target's RV spectrum can be directly extracted in the folded domain without the need to restore the signal to the original domain.

[0058] 2. Folded-domain RV spectrum extraction: By performing FFT operation within the folded domain, the target's RV spectrum is directly obtained, avoiding the complex signal reconstruction process.

[0059] 3. Folded-domain interference cancellation: Based on the radar processing result within the folded domain, the radar interference signal is reconstructed, and interference cancellation is performed within the folded domain. In the folded domain, the least mean square error (LMS) algorithm is used to eliminate the interference of strong sensing signals on weak communication signals. The reconstruction of the radar interference signal is:

[0060]

[0061] where g m is the mth channel coefficient, r is the amplitude part of the radar channel gain, Δf is the subcarrier spacing, l is the subcarrier index, Θ is the time delay, e -j2πlΔfΘ is the phase offset caused by the time delay, f d is the Doppler shift, μ is the configuration parameter of the OFDM subcarrier spacing, T is the sampling period,

[0062] The output of the folded-domain LMS algorithm is represented as:

[0063]

[0064] where y is the received signal, is the reconstructed radar interference signal.

[0065] 4. Algorithm principle:

[0066] 4.1 Folding operation algorithm

[0067] The received signal y is folded, i.e. the signal amplitude is constrained in the interval [-λ, λ] to avoid ADC saturation:

[0068]

[0069] where k is the kth time, also known as the sampling point index, and T is the sampling period.

[0070] 4.2 Folded-domain interference suppression algorithm

[0071] Step 1: Obtain the RV spectrum in the folded domain radar processing, extract target parameters and where, is the time delay in the range dimension of the RV spectrum, is the Doppler shift in the velocity dimension of the RV spectrum;

[0072] Step 2: Reconstruct the radar jamming signal according to the target parameters:

[0073] For the jamming reference reconstruction signal of the mth receiving channel at the kth moment:

[0074]

[0075] If processed by a single channel:

[0076] Step 3: Jamming cancellation: realize by LMS adaptive filtering in the folded domain:

[0077]

[0078] is the folded domain output of the SOI estimate value at moment k, μ s is the step size, the folded domain output is equivalent to the original domain SOI.

[0079] 5. Implementation method:

[0080] 5.1 Input and preprocessing

[0081] Receive signal y(t) → analog folding circuit → discrete folding sequence y[k].

[0082] Parameter setting: folding threshold λ, step size μ s , filter length L.

[0083] 5.2 Folded domain radar processing

[0084] Perform two-dimensional FFT on the folded signal to obtain the RV spectrum.

[0085] Extract the range and Doppler shift of the target.

[0086] 5.3 Jamming reconstruction

[0087] According to the RV spectrum parameters and the known transmitted waveform s rad [k], generate the jamming reference reconstruction signal

[0088] Construct the reference input vector x[k].

[0089] 5.4 Folded domain LMS jamming cancellation

[0090] The folded-domain LMS interference rejection framework is shown in Fig. 1. Figure 2

[0091] The ultimate goal is to recover the SOI from the received signal y[k] which contains strong interference. The received signal y[k] is folded by a factor of M λ λ (y(kT)) and a folding threshold λ > 0 as input to an adaptive filter (e.g. LMS) to estimate the strong interference component and subtract it from the received signal. The remaining signal is the recovered SOI, i.e. the target signal

[0092] The detailed steps are as follows:

[0093] Step 1: Initialization

[0094] Initialize the adaptive filter weight vector to zero w[0] = 0. The adaptive filter weight vector is a complex Lx1 vector, where L is the filter length, i.e. the filter order.

[0095] The filter input reference signal vector x[k] is composed of the interference reference reconstruction at the current time instant and the previous L-1 time instants, T denotes the transpose.

[0096] Step 2: Main loop (process each time instant k)

[0097] Ensure that the input fold: y λ λ (k) = M

[0098] Estimate the interference: Use the current adaptive filter weight w and the reference signal x[k] to estimate or predict the current interference signal H denotes the conjugate transpose. The adaptive filter weight w is updated by the step size μ and the previous target signal estimate, which can more and more accurately estimate the time-varying interference.

[0099] Folded-domain cancellation and SOI extraction: It consists of three sub-steps:

[0100] a) Folded interference estimation: first estimate the interference Also perform the same folding operation M λ ​​​​Because the subtraction must be done in the same folding domain, otherwise it will cause error due to the mismatch of modulo operation.

[0101] b) Cancellation: from the received signal y λ Subtract the folded interference estimate from [k]

[0102] c) Refolding: fold the cancellation result again M λ To ensure the final SOI estimate Also falls in the legal interval [-λ, λ] and completes the de-mixing.

[0103] Step 3: Output

[0104] Output the recovered signal of interest

[0105] All subtractions and reconstructions are done in the folding domain, circumventing the nonlinearity problem caused by signal folding / quantization, making the standard adaptive filtering theory applicable, suitable for scenarios where strong interference overwhelms weak targets. The folding domain computation architecture proposed in this invention enables the LMS algorithm to run equivalently in the folding domain and the original domain, maintaining mathematical consistency while reducing computational complexity. In radar, this can be a strong third-party jammer signal; in communication, this can be a strong signal in the adjacent channel.

[0106] Simulation experiment

[0107] To verify the effectiveness of the invention, simulation and hardware measurement are carried out:

[0108] System configuration: carrier frequency 26GHz, subcarrier number 512, subcarrier interval 120kHz, modulation method 16-QAM, target distance 4010m, target speed 50m / s, signal-to-noise ratio SNR=30dB.

[0109] Simulation results: when the folding threshold λ=1 / 4, the system dynamic range is improved by about 6dB, and the mean square error MSE is only degraded within 1.5dB.

[0110] Hardware verification: in the folding board measurement of Keysight vector signal source + Tektronix oscilloscope, the peak position of RV spectrum is completely consistent in the folding domain and the original domain. The constellation diagram shows: before cancellation, the error vector magnitude EVM of the communication signal is 51%, which cannot be normally demodulated; after folding domain cancellation, the EVM is reduced to 12%, which can correctly complete QPSK demodulation.

[0111] Multi-user scenario: in a four-user communication environment, when SINR>-15dB, the method of the invention makes the total communication rate at least 20% higher than the traditional method.

Claims

1. An uplink ISAC system interference suppression method based on signal folding, characterized in that, Includes the following steps: The received signal is folded to limit its amplitude within a preset folding threshold range; Perform a Fast Fourier Transform (FFT) operation on the target echo within the folded domain to extract the target's range-velocity (RV) spectrum. The interference reference reconstructed signal is obtained based on the RV spectrum; in, Let s be the interference reference reconstructed signal at time k. rad [k] represents the radar transmission signal at time k, m is the channel number, and g m Let be the m-th channel coefficient, Δf be the subcarrier spacing, l be the subcarrier index, μ be the configuration parameter for the OFDM subcarrier spacing, and T be the sampling period. For the time delay in the RV spectral distance dimension, Due to the Doppler frequency shift in the velocity dimension of the RV spectrum; An adaptive filtering algorithm is used within the folded domain to estimate and eliminate radar interference signals, thereby recovering the communication signal.

2. The method as described in claim 1, characterized in that, The folding operation is as follows: Among them, M λ Let λ be the folding operation function, x be the input signal for the folding operation, and λ be the folding threshold. This indicates rounding down to the nearest integer.

3. The method as described in claim 2, characterized in that, The adaptive filtering algorithm is used to estimate the radar interference signal within the folded domain, specifically as follows: in, For the radar interference signal estimated at time k, the input reference signal vector of the filter is x[k]. H This represents the conjugate transpose, reconstructed from the disturbance reference at time k and L-1 previous times. composition, T denoted as transpose, L is the length of the filter, and w is the adaptive filter weight.

4. The method as described in claim 3, characterized in that, The adaptive filtering algorithm used to eliminate radar interference signals within the folded domain is specifically as follows: Among them, y λ [k]=M λ (y[k]), where y[k] is the received signal that underwent a folding operation at time k.

5. The method as described in claim 4, characterized in that, The method as described in claim 2, characterized in that it includes the following sub-steps: a) Folded Interference Estimation: First, the estimated radar interference signal... Perform folding operation M λ The radar interference estimate after folding is obtained; b) Cancellation: In the received signal y after the folding operation... λ [k] minus the radar interference estimate after folding The result after elimination is obtained; c) Fold again: Fold the result after cancellation again. λ To ensure the final SOI estimate It also falls within the valid interval [-λ,λ] and completes the defuzzification.

6. The method as described in claim 5, characterized in that, The method for updating the adaptive filter weights is as follows: w[k+1] represents the filter weight at time k+1, μ s The step size is set.

Images