A training sequence-assisted interference detection method

Abstract

The present invention belongs to the field of wireless communication technologies, and particularly relates to an interference detection method assisted by training sequences. When there is a useful signal, by using the prior knowledge of the training sequences in the useful signal, through robust channel estimation under interference conditions, the useful signal is reconstructed and cancelled, so as to obtain a cleaner interference signal for further detection. Simulation verification shows that its interference detection performance is better than that of traditional detection schemes, and can approach the interference detection ability in the silent period.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communication, and particularly relates to a training sequence-assisted interference detection method. Background Art

[0002] Interference detection technology is one of the key technologies in the field of communication anti-interference. It can not only identify the presence or absence of interference signals, but also provide parameters such as the interference frequency point position, interference power, and interference type, providing prior information for the interference suppression and anti-interference technology of communication systems. Among the existing time-domain and various transform-domain interference detection technologies, the frequency-domain-based interference detection algorithm is the most widely used.

[0003] The energy detection algorithm does not require any form of assumption about the interference signal and the user signal, and is widely used because of its simple principle and easy implementation. However, the performance of the energy detection algorithm is affected by problems such as user useful signals, multipath fading, and hidden terminals.

[0004] For conventional interference detection to achieve better detection effects, it generally requires that the waveform design reserves a silent period without user useful signals or reduces the impact of signals on interference detection through spreading. This assumption requires that the applicable waveform needs to be specially customized, and the silent period will reduce the spectral efficiency, resulting in performance loss of the detection algorithm under low signal-to-interference ratio, limiting the application scope of the algorithm. Summary of the Invention

[0005] In view of the above problems, the present invention proposes a training sequence-assisted interference detection method. When there is a useful signal, using the prior knowledge of the training sequence in the useful signal, through robust channel estimation under interference conditions, the useful signal is reconstructed and canceled, so as to obtain a cleaner interference signal for further detection. Simulation verification shows that its interference detection performance is better than that of traditional detection schemes, and it can approach the interference detection ability under the silent period.

[0006] The technical solution of the present invention is as follows:

[0007] A training sequence-assisted interference detection method, as Figure 1 shown, includes the following steps:

[0008] S1. The receiving end extracts the signal of the training sequence segment in the received signal:

[0009] y(n) = h * d(n) + w(n) + j(n), n = 1,..., N

[0010] Among them, h represents the channel impulse response of the channel through which the signal passes, d(n) represents the training sequence, w(n) represents the noise sequence, j(n) represents the interference sequence, and N represents the length of the training sequence. For the scenario where the receiver knows the training sequence, for the scenario where the receiver knows the training sequence, to cancel the training sequence, an estimated value of h needs to be obtained first;

[0011] S2. Use the training sequence to perform channel estimation to obtain the estimated value of the channel impulse response Specifically:

[0012] Perform DFT transformation on y(n) to obtain:

[0013] Y(n) = Fh * Fd(n) + Fw(n) + Fj(n)

[0014] = diag(Fd(n))Fh + Fw(n) + Fj(n)

[0015] Among them, Y(n) is the measured value, F is the DFT matrix of size N×N, and diag(Fd(n))F = Ψ is the sensing matrix;

[0016] Initialize the residual r (0) = Y(n), the index set The loop count t = 1, input the sparsity K of the channel response and the sensing matrix Ψ' of size N×N:

[0017]

[0018] Among them is the submatrix composed of the first N τ columns of the matrix F, and N τ is the maximum multipath delay;

[0019] Construct an orthogonal index set to make the sensing matrix satisfy orthogonality:

[0020]

[0021] Among them, pos is used as the orthogonal index set, i is the initial index value, and 2 m is the interval value, and 2 k is the total number of non-zero values in pos, satisfying 2 k ≥ N τ ;

[0022] Take the submatrix of the sensing matrix selected according to the index value as the sensing matrix:

[0023] Ψ = diag(pos)Ψ'

[0024] Calculate the correlation matrix:

[0025] Λ (t-1) = ΨH r (t-1)

[0026] Find the row index with the largest correlation value in the correlation matrix:

[0027]

[0028] Add the found row index to the index set:

[0029] S (t) = S (t-1) ∪{i}

[0030] The 2 (t) ×t matrix composed of the columns of Ψ corresponding to the indices in S k forms Ψ (t) ;

[0031] Reconstruct the signal using the least squares method:

[0032] h (t) = (Ψ (t)H Ψ (t) ) -1 Ψ (t)H y

[0033] Update the current residual:

[0034] r (t) = y - Ψ (t) h (t)

[0035] Let t = t + 1, and determine whether t is greater than k. If not, continue to calculate the correlation matrix for iteration; otherwise, stop the iteration and output:

[0036]

[0037] S3. Obtain the estimated value of the channel time-domain impulse response After that, perform Fourier transform on it to obtain the frequency-domain expression of the channel impulse response For the scenario where the training sequence is known at the receiving end, the received sequence separates the influence of the useful signal in the time-frequency domain, and the remaining sequence only contains the sequence of noise and interference

[0038]

[0039] S4. Perform frequency-domain interference detection on the sequence after removing the influence of the training sequence ;

[0040] S5. Output the interference detection result

[0041] The beneficial effects of the present invention are as follows. For interference detection in the presence of useful signals, a training sequence-assisted interference detection method is proposed. By performing channel estimation, based on the known signals, the useful signals passing through the channel in the received signals affected by interference are eliminated, obtaining a signal similar to the silent period signal, thereby improving the interference detection performance. Through simulation verification, the training sequence-assisted interference detection method can effectively improve the interference detection performance under signal conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic diagram of the logical flow of the system of the present invention.

[0043] Figure 2 It is a simulation diagram of the interference detection performance based on channel estimation in an AWGN channel.

[0044] Figure 3 It is a simulation diagram of the interference detection performance based on channel estimation in a fading channel. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The present invention will be further described in detail below with reference to the drawings and simulations.

[0046] First, when performing interference detection under the condition of the presence of useful signals, for the scenario where the receiver knows the training sequence, after obtaining the estimated value of the channel impulse response, the received sequence can separate the influence of the transmitted signal, and the remaining sequence only contains the information of noise and interference. That is, the purpose of eliminating the useful signal for interference detection in the period similar to the silent period is achieved. However, since interference will affect the performance of channel estimation, a robust channel estimation needs to be adopted. The present invention uses the Orthogonal Matching Pursuit (OMP) algorithm, which is a classic algorithm in the field of compressive sensing, to estimate the channel impulse response under interference conditions. This algorithm has the characteristics of simplicity and high efficiency.

[0047] To illustrate the superiority of this scheme in interference detection performance, the beneficial effects of the present invention will be verified through simulation below.

[0048] First, simulations are carried out in an AWGN channel to verify the theoretical correctness of the present scheme. Combining Monte Carlo simulations, Figure 2 reveals the detection rate performance of interference detection based on a training sequence with a length of 1024, and the detection rates in the presence of signals, the silent period, and the detection assisted by using the known training sequence are respectively simulated. According to Figure 2 the results, the detection rates in the silent period and the detection assisted by using the known training sequence are basically coincident, and there is a performance gain of about 12 dB compared with the interference detection in the presence of signals at SNR = 10 dB.

[0049] Figure 3Reveals the detection rate performance of interference detection based on a training sequence of length 1024 in a fading channel. The detection rates in the presence of signals, silent periods, and detection assisted by known training sequences are simulated respectively. For detection assisted by training sequences, the detection rate curves are simulated with interval values of 16, 64, 256, and under ideal channel estimation. According to Figure 3 the results, the detection rate of detection assisted by training sequences increases as the interval value decreases, and as the interval value decreases, the detection rate of detection assisted by training sequences gradually approaches the detection rate in the silent period. At the same time, the detection rate under ideal channel estimation has a performance gain of about 12 dB compared to signal interference detection at SNR = 10 dB.

Claims

1. A training sequence-assisted interference detection method, characterized in that, Including the following steps: S1. The receiving end extracts the signal of the training sequence segment in the received signal: ， Among them, represents the channel impulse response of the channel through which the signal passes, represents the training sequence, represents the noise sequence, represents the interference sequence, represents the length of the training sequence. For the scenario where the receiver knows the training sequence, to cancel the training sequence, it is first necessary to obtain the estimated value; S2. Use the training sequence to perform channel estimation to obtain the estimated value of the channel time-domain impulse response , specifically: Pair Performing DFT transformation gives: ， Among them, is the measured value, is the DFT matrix of size , is the sensing matrix; Initial residual , index set , loop count , sparsity of the input channel response and sensing matrix of size : : ， wherein is a sub-matrix formed by the first columns of the matrix ; is the maximum multipath delay; Construct an orthogonal index set to make the sensing matrix satisfy orthogonality: ， wherein is an orthogonal index set, is the initial index value, is the interval value, is the total number of non-zero values in ; Use the sub-matrix of the sensing matrix selected according to the index value as the sensing matrix: ， Calculate the correlation matrix: ， Find the row index with the largest correlation value in the correlation matrix: ， Add the found row index to the index set: ， composed of the columns corresponding to the indices in in forming a matrix ; Reconstruct the signal using the least squares method: ， Update the current residual: ， Let , and determine is greater than . If not satisfied, continue to calculate the relevant matrix for iteration; otherwise, stop the iteration and output: ， S3. Obtain the estimated value of the channel time-domain impulse response After that, perform Fourier transform on it to obtain the frequency-domain expression of the channel impulse response For the scenario where the receiver knows the training sequence, the received sequence separates the influence of the useful signal in the time-frequency domain, and the remaining sequence only contains the sequence of noise and interference : ， S4. Perform frequency-domain interference detection on the sequence after removing the influence of the training sequence ; S5. Output the interference detection result.

Images