A Joint Estimation Method for Roll-off Factor and Order of a Communication Signal Shaping Filter

Abstract

The present invention belongs to the field of electronic countermeasures, and particularly relates to a method for jointly estimating the roll-off coefficient and order of a communication signal shaping filter. This method utilizes the characteristic that the mismatch between the roll-off coefficient and order in the modulator and demodulator has little influence on the bit error performance of the M-PSK signal. A fixed roll-off coefficient and order are selected from the common parameters of the shaping filter to demodulate the communication signal at the receiving end, and the symbol information of the communication signal is obtained. Let a roll-off coefficient range vector alpha with a length of N and a filter order vector S with a length of M be set. An NxM shaping filter parameter matrix is composed of these two vectors, and the symbol information of the communication signal is modulated by the shaping filter determined by all the parameters in this matrix. The mean square error is calculated by comparing the modulated signals with the signals after the lower sideband of the received signal, and a mean square error matrix is obtained. The parameters corresponding to the minimum mean square error are the joint estimation results of the roll-off coefficient and order of the shaping filter.

Description

Technical Field

[0001] The present invention belongs to the field of electronic countermeasures, and particularly relates to a method for jointly estimating the roll-off coefficient and order of a communication signal shaping filter. Background Art

[0002] In modern digital communication, the spectrum range of baseband signals is relatively wide. In order to enable better signal transmission in the channel, it is necessary to band-limit the signal through a shaping filter at the transmitting end, which will introduce inter-symbol interference. Using a matched filter at the receiving end can overcome this inter-symbol interference, thereby achieving the best receiving effect.

[0003] The Nyquist criterion states that any filter as long as its impulse response satisfies

[0004]

[0005] can eliminate inter-symbol interference. A filter that satisfies the Nyquist criterion is called a Nyquist filter. When the modulated signal is transmitted in the channel, distortion will be introduced. We can use an equalizer with a transfer function opposite to that of the channel to completely eliminate the distortion. Then the entire transfer function h(t) can be approximated as the product of the transmitter and receiver filter functions. An effective end-to-end transfer function, h(t) is often implemented using filters with a transfer function of at both the receiver and transmitter ends.

[0006] In practical applications, the shaping filter and the matched filter often adopt the form of a square root raised cosine roll-off filter. The transfer function of the square root raised cosine roll-off filter is shown in Equation (2)

[0007]

[0008] where α is called the roll-off coefficient, T is the symbol period, and the bandwidth B of the filter is B=(1 + α) / T. Its two important parameters are the roll-off coefficient and the filter order. The roll-off coefficient describes the bandwidth characteristics of the filter. It can be proved that when the roll-off factors and orders in the shaping filter and the matched filter are the same, the system performance is optimal. This is easily satisfied in cooperative communication, but in non-cooperative communication, the receiver cannot know in advance the roll-off coefficient and order selected by the transmitting end. Therefore, when the receiving end performs signal processing, either a fixed value is adopted for the roll-off coefficient and order of the shaping filter to perform the same processing on all received signals; or the roll-off coefficient and order of the received signal are estimated and then the received signal is processed.

[0009] In fact, the mismatch between the roll-off factor and the order in the modulator and demodulator has little effect on the bit error performance of the M-PSK signal. That is, at the receiving end, only a suitable roll-off factor and order need to be assumed to demodulate the signal. Therefore, there is little research on the estimation of the roll-off factor and order of the received signal. However, in some specific scenarios, when the receiving party receives the target signal, it will be interfered by the communication signal, making the received signal an aliased signal of the communication signal and the target signal. Usually, the interfering communication signal is a large signal and the target signal is a small signal. Assuming that the method of reconstruction and cancellation is used to separate the target signal, it is necessary to estimate the parameters of the communication signal and then reconstruct the communication signal, and it is necessary to accurately estimate the roll-off factor and order of the shaping filter of the communication signal. Summary of the Invention

[0010] The object of the present invention is to provide a joint estimation method for the roll-off factor and order of a communication signal shaping filter in view of the above problems. This method utilizes the characteristic that the mismatch between the roll-off factor and the order in the modulator and demodulator has little effect on the bit error performance of the M-PSK signal, demodulates and reconstructs the aliased signal, and finds the minimum mean square error between the time-domain waveforms of the reconstructed signal and the aliased signal to complete the joint estimation of the roll-off factor and order of the shaping filter.

[0011] The technical solution adopted by the present invention is as follows:

[0012] A joint estimation method for the roll-off factor and order of a communication signal shaping filter first sets a roll-off factor range vector alpha with a length of N and a filter order vector S with a length of M. These two vectors form an NxM shaping filter parameter matrix. First, parameter estimation is performed on the received aliased signal, including modulation parameters such as carrier frequency and code rate. Utilizing the characteristic that the mismatch between the roll-off factor and the order in the modulator and demodulator has little effect on the bit error performance of the M-PSK signal, a fixed roll-off factor and order are selected from the common parameters of the shaping filter to demodulate the communication signal at the receiving end to obtain the symbol information of the communication signal. The symbol information of the communication signal is modulated using the shaping filter determined by all the parameters in the parameter matrix, and the time-domain mean square error is calculated between the modulated signals and the signal after the received signal is down-converted to obtain a mean square error matrix. The parameters corresponding to the minimum mean square error are the joint estimation results of the roll-off factor and order of the shaping filter. Figure 1 The flow diagram of this technical solution is shown as follows, including the following steps:

[0013] S1. Set a roll-off factor range vector alpha with a length of N and a filter order vector S with a length of M. These two vectors form an NxM shaping filter parameter matrix.

[0014] S2. Estimate the parameters of the received signal to obtain information such as the carrier frequency and code rate of the BPSK signal modulation parameters.

[0015] S3. Arbitrarily select a set of data from the filter parameter matrix, and jointly demodulate the received signal with the modulation parameters obtained in S2 to obtain the signal S after the received signal is down-converted. 1 (t) and the symbol information of the communication signal.

[0016] S4. Then, modulate the obtained communication signal symbols with the same modulation parameters estimated in S2, design a shaping filter with the parameters in the shaping filter parameter matrix, and obtain the reconstructed BPSK signal under different filter parameters.

[0017] S5. Compare the mean square error of the time-domain waveforms of different reconstructed signals and the signal S after the received signal is down-converted. 1 (t). The parameters corresponding to the minimum mean square error are the joint estimation results of the roll-off factor and order of the shaping filter.

[0018] The present invention has the following beneficial effects. The present invention utilizes the characteristic that the mismatch between the roll-off factor and order in the modulator and demodulator has little effect on the bit error performance of the M-PSK signal to jointly estimate the roll-off factor and order of the shaping filter, and can obtain a relatively accurate estimation result without too many prior conditions. The method is simple and has good effects. Description of the Drawings

[0019] Figure 1 Flowchart of the technical solution;

[0020] Figure 2 Bit error rate of BPSK signal demodulation under different filter parameters of the receiver;

[0021] Figure 3 Estimation accuracy rate of the roll-off factor of the BPSK signal in the frequency domain and time domain;

[0022] Figure 4 Estimation accuracy rate of the order of the BPSK signal filter in the frequency domain and time domain;

[0023] Figure 5 Bit error rate of BPSK+LFM signal demodulation;

[0024] Figure 6 Estimation accuracy rate of the roll-off factor of the filter in the aliased signal;

[0025] Figure 7 Estimation accuracy rate of the order of the filter in the aliased signal. Detailed Embodiment

[0026] The technical solution of the present invention will be further described below in conjunction with the drawings and simulations.

[0027] In this example, the proposed method will be simulated and verified: Define the communication signal as S 1 (t), and the target signal as S 2 (t). The code rate of the communication signal is R s , and the sampling rate is F s . The communication signal is oversampled by F s / R s .

[0028] First, design a roll-off factor range vector alpha with a range of 0.2 - 0.6, an interval of 0.05, and a length of 9, and a filter order vector S with a range of 2 - 10 and a length of 9. These two vectors form a 9x9 shaping filter parameter matrix.

[0029] First, when only the communication signal exists, assume the communication signal is a BPSK signal. The received signal model at the receiver can be written in the form of Equation (3)

[0030]

[0031] where a i , T are the symbol values and symbol intervals of the communication signal respectively, and g T (t) is the impulse response of the root-raised cosine filter. Assume the number of signal symbols is 15000, the code rate is 15 MHz, the carrier frequency is 30 MHz, the sampling rate is 60 MHz, the oversampling is 4, the symbol truncation length N = 8, the roll-off factor α = 0.35, the number of Monte Carlo runs is 1000, the signal-to-noise ratio is set from 0 - 50 dB. Assume that except for the shaping filter parameters, other modulation parameters of the communication signal are estimated correctly, and only the shaping filter parameters have errors.

[0032] Then, based on the above, add an LFM signal as the target signal. The received signal model at the receiver can be written in the form of Equation (4)

[0033]

[0034] S 2 '(t) is the multipath of the target signal, N(t) represents the number of multipath signals generated by the reflection of the target signal, A i (t) represents the gain of the i-th signal received by the receiver, τ i (t) represents the delay generated by the i-th path, represents the Doppler frequency shift generated by the i-th path.

[0035] The modulation parameters of the BPSK signal remain unchanged. The bandwidth of the LFM signal is 5 MHz, the pulse width is 10 μs, the duty cycle is 0.2, the number of pulses is 20, the carrier frequency is the same as that of the communication signal, both are 30 MHz, the number of multipaths of the LFM signal is 4, the multipath channel conforms to the first Fresnel zone modeling, the communication signal is used as the interference signal, and the interference-to-signal ratio is set to 0 - 20 dB.

[0036] The simulation results are shown in the figure:

[0037] Figure 2 For the bit error rate of BPSK signal demodulation under different filter parameters of the receiver, it can be analyzed that the influence of the mismatch between the roll-off factor and the order in the modulator and demodulator on the bit error performance of the M-PSK signal is very small and can be ignored. Figure 3 and Figure 4 For the correct estimation rate of the roll-off factor and the order of the BPSK signal frequency-domain and time-domain filters, it can be seen from the figure that the correct rate of using the time-domain waveform to find the minimum mean square error to estimate the filter parameters is significantly higher than that of the frequency-domain comparison.

[0038] Figure 5 For the bit error rate of BPSK and LFM aliased signal demodulation under different filter parameters of the receiver, it can be analyzed that in the high interference-to-signal ratio environment, the bit error rate obtained by directly demodulating the aliased signal is consistent with the theoretical bit error rate of BPSK demodulation, that is, in the high interference-to-signal ratio environment, accurate BPSK signal symbols can be obtained by directly demodulating the aliased signal. Figure 6 and Figure 7 For the correct estimation rate of the roll-off factor and the order of the BPSK and LFM aliased signal time-domain filters, it can be seen from the figure that using this method can accurately estimate the shaping filter parameters at a signal-to-noise ratio of 15 dB, and the correct estimation rate increases with the increase of the interference-to-signal ratio.

Claims

1. A joint estimation method for the roll-off coefficient and order of a communication signal shaping filter, characterized in that, it includes the following steps: S1. Define a roll-off coefficient range vector alpha with a length of N and a filter order vector S with a length of M, and form a shaping filter parameter matrix of NxM from these two vectors; S2. Perform parameter estimation on the received signal to obtain the modulation parameters of the BPSK signal; S3. Arbitrarily select a set of data from the shaping filter parameter matrix defined in S1, and jointly demodulate the received signal with the modulation parameters obtained in S2 to obtain the signal S 1 (t) and the symbol information of the communication signal; S4. Modulate the symbol information of the obtained communication signal with the modulation parameters estimated in S2, design a shaping filter with the parameters in the shaping filter parameter matrix, and obtain the reconstructed BPSK signal under different filter parameters; S5. Compare the signals S after down-converting the different reconstructed signals and the received signal, and the mean square error under the time-domain waveform of 1 (t). The parameters corresponding to the minimum mean square error are the combined estimation results of the roll-off factor and the order of the shaping filter. 1 (t). The mean square error under the time-domain waveform, and the parameters corresponding to the minimum mean square error are the combined estimation results of the roll-off factor and the order of the shaping filter.

Images