CN105635021A - Pulse noise combined inhibition method in multicarrier communication system - Google Patents

Abstract

The invention belongs to the field of wireless communication, and relates to a joint suppression method of impulse noise in a multi-carrier communication system. Firstly, the signal to be processed z(k) is subjected to time-domain nonlinear preprocessing to obtain a y(k) signal; the y(k) The signal is FFT transformed into the frequency domain to obtain the frequency domain signal Y(k); then hard decision processing is performed in the frequency domain, and the hard decision signal Z'(k) is subjected to IFFT transformation to obtain the time domain signal y'(k); then use The signal y(k) is subtracted from the yˊ(k) signal to obtain nˊ(k), and the signal nˊ(k) is used to calculate the negative impulse noise estimate iˊ(k); finally, the negative impulse noise estimate iˊ( k) signal. The present invention combines nonlinear processing in the time domain and a feedback noise estimation method in the frequency domain to obtain a joint suppression method; and in Bernoulli-Gaussian and Middleton? Two typical impulse noise models of class A have passed the simulation verification. The results show that this method can greatly reduce the symbol error rate, and achieve faster speed and higher accuracy under the condition of high impulse noise power. suppression.

Description

A Joint Suppression Method of Impulse Noise in Multi-carrier Communication System

technical field

The invention belongs to the field of wireless communication, and relates to a pulse noise suppression technology in a multi-carrier communication system, in particular to a joint suppression method for pulse noise in a multi-carrier communication system.

Background technique

With the rapid development of wireless communication technology, the performance requirements for communication data transmission are getting higher and higher. On the other hand, wireless communication in the VHF/UHF frequency band is suffering from the interference of impulse noise. In addition to additive white Gaussian noise, impulse noise is another additive noise source that degrades the performance of wireless communication systems. It is mainly caused by spark plug ignition devices in automobiles, electrical switching equipment, high-voltage lines, and lightning. Compared with Gaussian white noise, the generation of impulse noise has the characteristics of suddenness, short pulse, high energy and sharpness. Studies have shown that the impact of impulse noise on wireless communication systems is mainly distributed in the VHF and lower UHF frequency bands. With the modernization of cities and the popularization of electrical equipment, the impact of impulse noise in this frequency range will be more serious.

In order to reduce the equalization complexity under the frequency selective fading channel and improve the robustness to impulse noise, OFDM (Orthogonal Frequency Division Multiplexing) has been proposed and has been widely used in such as DTMB (Digital Television Terrestrial Broadcasting), DVB (Digital Video Broadcasting), WiMAX (Global Microwave Interconnection Access) and other wireless systems and PLC power line carrier communication. It is mentioned in the literature that compared with single-carrier systems, OFDM is inherently robust to low-energy impulse noise, because the DFT (discrete Fourier transform) processing at the receiver side of OFDM converts an OFDM in the time domain Impulse noise within a symbol period extends to the frequency domain. On the one hand, because the impulse noise energy is averaged to all subcarriers, this weakens the impact of impulse noise to a certain extent, but on the other hand, because of the energy expansion effect of DFT, the high-energy impulse noise after frequency domain expansion This deteriorates the carrier-to-noise ratio of each subcarrier, which leads to more severe symbol decision errors.

Since the occurrence probability of impulse noise is small, and the impact of OFDM on impulse noise depends on the total energy of the impulse in a single OFDM symbol period, the OFDM system with a large number of carriers has more advantages in reducing the impact of impulse. For the wireless communication system in the low UHF frequency band, due to the relatively insufficient code interleaving depth and the number of carriers, additional algorithms must be used to suppress impulse noise at the physical layer.

In view of the ubiquitous impulse noise in the VHF/UHF frequency band, the current main suppression methods are time-domain nonlinear processing, increasing interleaving depth, filter compensation method, etc. Time-domain nonlinear processing mainly includes methods such as zeroing, clipping, and the combination of zeroing and clipping. These algorithms are simple, easy to implement, and show good characteristics in practical applications, especially for high-energy impulse noise. However, it is not easy to find a suitable threshold value for clearing or clipping in time-domain processing. At the same time, the time-domain processing algorithm itself will introduce negative noise, so this method cannot bring great performance improvements. Utilizing OFDM characteristics and adopting block interleaving technology in the time domain can further weaken the influence of impulse noise; this method shows good performance against different degrees of impulse noise under low-order modulation methods, but under high-order modulation , especially in a strong pulse environment, increasing the interleaving depth will only lead to a more serious bit error rate, and byte interleaving will also increase the complexity of the system; at the same time, this method has a certain dependence on the actual noise environment , but also need to make a trade-off between system performance and complexity. The filter compensation method is a periodic impulse noise suppression algorithm. This algorithm detects the impulse noise of multiple OFDM symbols in the frequency domain, and determines the position of the impulse noise in the subcarrier by the method of majority voting; at the same time, it designs a multi-order Adaptive IIR notch filter to suppress periodic noise, and then compensate the distorted received signal through IIR notch filter, and perform forward error correction (FEC) processing at the decoding end to obtain large bit errors at a certain bandwidth cost The efficiency has been improved; but the limitation of this method is also obvious, it can only suppress the periodic impulse noise, and has no suppression effect on the sudden impulse noise.

From the background above, it can be seen that in the actual noise environment, traditional impulse noise suppression methods face problems such as high system complexity, limited performance improvement space, and high dependence on the environment, making it difficult for the suppression performance to meet the requirements of high-speed and accurate data flow. Because impulse noise has the characteristics of suddenness, short duration, high energy, and sharpness, the harsh environment of impulse noise and the consideration of system complexity have high requirements for us to design suppression methods at the physical layer.

Contents of the invention

The object of the present invention is to provide a joint suppression method of impulse noise in a multi-carrier communication system, which combines the characteristics of impulse noise in the burst noise environment, and combines the nonlinear processing of the time domain with the feedback noise estimation method of the frequency domain; And it has passed the simulation verification under two typical impulse noise models of Bernoulli-Gaussian and MiddletonA. The results show that this method can greatly reduce the symbol error rate and achieve faster speed under the condition of high impulse noise power. Faster, more accurate suppression. The technical scheme adopted in the present invention is:

A joint suppression method of impulse noise in a multi-carrier communication system, comprising the following steps:

Step 1, performing time-domain nonlinear preprocessing on the signal z(k) to be processed to obtain a y(k) signal;

Step 2, FFT transforming the y(k) signal through the time domain nonlinear preprocessing into the frequency domain to obtain the frequency domain signal Y(k);

Step 3, performing hard decision processing in the frequency domain to obtain the signal Z'(k);

Step 4, performing IFFT transformation on the signal Z'(k) after the hard decision to obtain the time domain signal y'(k);

Step 5, obtain signal y(k) with step 1 and subtract step 4 to obtain y'(k) signal and obtain n'(k);

Step 6, using the signal n'(k) to calculate the negative impulse noise estimate i'(k);

Step 7. Subtract the negative impulse noise estimation i'(k) signal obtained in step 6 from y(k) obtained in step 1, and obtain the signal z'(k) after joint suppression of impulse noise.

Further, the time-domain nonlinear preprocessing in the step 1 adopts a zero-setting method.

The concrete steps of described step 6 are:

First, calculate the average power of the signal n'(k) as S,

$\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; =</span>\sqrt{\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

The negative impulse noise is then estimated as:

$i^{\&prime;}\left(k\right)=\{\begin{array}{cc}{\mathrm{no}}^{\&prime;}\left(k\right)& \left|{\mathrm{no}}^{\&prime;}\left(k\right)\right|\&Greater\; Equal;C\&Center\; Dot;S\\ 0& \left|{\mathrm{no}}^{\&prime;}\left(k\right)\right|<C\&CenterDot;S\end{array},$ C is the threshold value of impulse noise estimation.

The effect of the present invention is to provide a joint suppression method of impulse noise in a multi-carrier communication system, which combines the characteristics of impulse noise in the burst noise environment and combines the nonlinear processing in the time domain with the feedback noise estimation method in the frequency domain. Combination; effectively suppress the impact of impulse noise on multi-carrier communication, greatly improving the robustness of the wireless communication system.

Description of drawings

Figure 1 is a flow chart of the system.

Fig. 2 is a flow chart of time-frequency joint suppression.

Figure 3 shows the sampled values of Bernoulli-Gaussian noise.

Figure 4 is the sampling value of MiddletonA type noise.

Fig. 5 is a simulation result diagram of the optimal threshold of Blanking.

Fig. 6 is a theoretical performance curve diagram of different Blanking threshold values.

Fig. 7 is a schematic diagram of a hard decision process.

Fig. 8 is the elimination process in joint suppression of impulse noise.

Figure 9 shows the performance curves of different suppression methods from the perspective of SNR under the Bernoulli-Gaussian noise model.

Figure 10 shows the performance curves of different suppression methods from the perspective of SINR under the Bernoulli-Gaussian noise model.

Fig. 11 is the performance curve of the joint suppression algorithm under the Bernoulli-Gaussian noise model.

Figure 12 shows the performance curves of different suppression methods from the perspective of SNR under the Middleton noise model.

Figure 13 shows the performance curves of different suppression methods from the perspective of SINR under the Middleton noise model.

detailed description

The present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments.

The system flow chart of this embodiment is shown in Figure 1, which shows all the modules through which the data flow passes, which is also the functional block diagram of the OFDM system standard. In order to reproduce the impulse noise, two types of Bernoulli-Gaussian and MiddletonA are added to the channel A typical impulse-noise model assuming ideal channel estimation and ideal timing synchronization.

The implementation block diagram of time-frequency joint suppression is shown in Figure 2, which is mainly composed of time-domain nonlinear processing module, FFI and IFFT modules, hard decision device and impulse noise estimator. z(k) represents the signal to be suppressed received by the receiving end, y(k) represents the signal that has undergone time-domain nonlinear preprocessing, Y(k) represents the result of FFT after the preprocessed signal, Z'(k) is the result of the Y(k) signal passing through the hard decision device, and y'(k) is the result of the IFFT of the hard decision signal back to the time domain. The estimated value n'(k), i'(k) represents the result of the impulsive noise estimator, and finally subtract the impulsive noise estimated value i'(k) from y(k) to obtain z'(k).

First, two types of impulse noise models, Bernoulli-Gaussian and MiddletonA, are introduced. These two types of noise are additive noise. The Bernoulli-Gaussian impulse noise model is a Gaussian noise model, in which the impulse noise is distributed in all time-domain symbols, expressed as the product of an independent and identically distributed Bernoulli random process and a zero-mean Gaussian process, and its main parameters There are: the occurrence probability P of burst noise, and the variance of additive Gaussian white noise and the variance of burst noise

Under the Bernoulli-Gaussian noise model, the transmitted data is received with probability P and power The interference of the noise, otherwise it is only affected by the variance of Gaussian white noise interference, so its probability density function is expressed as:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; P</span>}}{\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; g</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}}\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; \&lsqb;</span><span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; g</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; \&rsqb;</span><span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; P</span>}}{\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; \&pi;\&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; g</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; g</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span>$

The MATLAB simulation time domain sequence is shown in Figure 3.

Middleton is a non-Gaussian narrow-band noise model whose main parameter is Gaussian white noise Impulse noise Its probability density function is expressed as:

$\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; A</span>}}\underset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; \&infin;</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}\frac{{\mathrm{<span\; class="notranslate">\; A</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}}{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; !</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; \&pi;\&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}{<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span>$

in,

${\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; A</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Gamma;</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Gamma;</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; \&Gamma;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; G</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; /</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; I</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; G</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; I</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}$

The probability density function of type A noise comes from the assumption that the number of pulses affecting the receiver obeys the Poisson distribution, and the parameters A and Γ control the pulse of the noise. A is the pulse index, indicating the average number of pulses that affect the normal demodulation of the receiver in one symbol duration, Γ is the power ratio of pulse noise and Gaussian noise; when A<1, the pulse of the noise becomes stronger, When A>1, the distribution of noise is close to Gaussian distribution; when Γ<1, the pulse performance is strong, and when Γ>1, the Gaussian feature is more obvious; the simulation time domain sequence is shown in Figure 4.

The time-domain nonlinear processing module adopts the method of zeroing (Blanking), and the impulse noise exceeding a specific threshold T is directly cleared at this time point:

${\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; \{</span>\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; T</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; T</span>}\end{array}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}$

Among them, N is the number of subcarriers, and T is the threshold value. There are two methods for its selection: first, through simulation, we know that when the threshold value is too high, many impulse noises may not be detected. The probability of missed detection will increase. When the threshold value is too low, the transmitted signal data may be treated as impulse noise, which will increase the probability of false alarms; under the simulation results of a large amount of data, there must be an optimal gate The limit value makes under the threshold value of this point, the false alarm probability and the missed detection probability reach the optimum, and the symbol bit error rate of the system is the lowest at this moment; Signal power divided by the sum of impulse noise and background noise power), the relationship between the threshold value and the bit error rate (using 16QAM modulation method), it can be seen from the figure that the position of the optimal threshold value appears at T = about 3.1. Second, obtained through theoretical derivation, the output maximum signal-to-noise ratio formula in Blanking processing:

$\mathrm{<span\; class="notranslate">\; \&gamma;</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; \&lsqb;</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}{<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&rsqb;</span>}{{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}$

in,

${\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\underset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\frac{{\mathrm{<span\; class="notranslate">\; T</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; e</span>}\frac{{\mathrm{<span\; class="notranslate">\; T</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

$\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; \&lsqb;</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}{<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&rsqb;</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; \&lsqb;</span>{\mathrm{<span\; class="notranslate">\; T</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&rsqb;</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\frac{{\mathrm{<span\; class="notranslate">\; T</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}<span\; class="notranslate">\; )</span>$

It is not difficult to draw the curves of output SNR and SIR (signal-to-interference ratio, the ratio of signal power to impulse noise power) under different threshold values T through the above formula, as shown in Figure 6. It can be seen from the figure that different power Under impulse noise, T=3.1 is a compromise value for outputting high SNR.

The FFT and IFFT modules are also the most basic modules in the OFDM system. FFT is a fast algorithm for realizing DFT (Discrete Fourier Transform). The transformation formulas of DFT and IDFT are as follows:

$\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; j</span>}\frac{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}}$

$\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}\frac{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}}$

Hard decision is also a nonlinear processing process. This module is based on the constellation mapping of the OFDM system; the hard decision device can help the system compensate for the negative noise introduced during time-domain nonlinear processing. The hard decision process of 16QAM in this embodiment is as follows: Figure 7 shows.

After FFT, hard decision, IFFT and other processes, the negative impulse noise power introduced by time-domain nonlinear processing has been equally divided into each subcarrier, and subtracted from y(k) to obtain the estimated value of impulse noise and background noise; Because negative noise is introduced by time-domain nonlinear preprocessing, the power of negative noise will not be very large. After FFT transformation, the power of negative impulse noise is equally divided into each subcarrier. The judgment process will not cause large symbol judgment errors; the existence of negative impulse noise only slightly raises the noise plane of y'(k), and the subtraction of y(k) and y'(k) actually extracts y(k) The useful data components in , which theoretically illustrate the feasibility of estimating n'(k).

The input signal n'(k) of the impulse noise estimator is the result of subtracting the hard decision feedback signal y'(k) from the non-linear preprocessed signal y(k), and y'(k) actually estimates the time Negative impulse noise and background noise introduced by domain nonlinear preprocessing. Since the negative impulse noise energy is more than ten dB higher than the background noise energy, the negative impulse noise can be detected relatively easily in the background noise, and the average power of the n'(k) signal is calculated as S,

$\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; =</span>\sqrt{\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Just set the impulse noise estimation threshold value C, then the estimation of negative impulse noise is:

${\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>& <span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; S</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; S</span>}\end{array}\right.$

Finally, negative impulse noise is eliminated in the time domain:

z'(k)=y(k)-i'(k),k=0,...,N-1

The elimination process of impulse noise in joint suppression is shown in Figure 8.

The OFDM simulation platform will be constructed below, the number of subcarriers N=4096, 16QAM modulation mode, non-linear preprocessing zero-setting threshold T=3.1, impulse noise estimation threshold C=4.5, the signal is divided into I and Q two-way transmission, and finally given out the simulation results.

The specific steps of the joint suppression method of impulse noise are as follows:

Step 1. Perform zero-setting nonlinear preprocessing on the signal z(k) to be processed, clear the sampling points greater than the threshold value T, and not process the sampling points lower than the threshold value, and obtain the y(k) signal;

Step 2, FFT transforming the y(k) signal through the time domain nonlinear preprocessing into the frequency domain to obtain the frequency domain signal Y(k);

Step 3, performing hard decision processing in the frequency domain to obtain the signal Z'(k);

Step 4, performing IFFT transformation on the signal after the hard decision to obtain the time domain signal y'(k);

Step 5, obtain signal y(k) with step 1 and subtract step 4 to obtain y'(k) signal and obtain n'(k);

Step 6, using the signal n'(k) to calculate the negative impulse noise estimate i'(k);

Step 7: Subtract the negative impulse noise estimate i'(k) signal obtained in step 1 from y(k) obtained in step 1 to obtain the final impulse noise joint suppressed signal z'(k).

Figures 9 to 11 show the simulation performance of the present invention under the Bernoulli-Gaussion noise model, and Figures 12 and 13 show the simulation performance under the Middleton noise model. Through a series of simulations on different background noise powers, different impulse noise powers and two noise models, the simulation results show that the method of time domain Blanking nonlinear processing combined with hard decision feedback reprocessing has superior suppression performance, especially The suppression performance of impulse noise of different powers is very good, very close to the theoretical curve.

The above is only a specific embodiment of the present invention. Any feature disclosed in this specification, unless specifically stated, can be replaced by other equivalent or alternative features with similar purposes; all the disclosed features, or All method or process steps may be combined in any way, except for mutually exclusive features and/or steps.

Claims (3)

1. A joint suppression method of impulse noise in a multi-carrier communication system, comprising the following steps: Step 1, performing time-domain nonlinear preprocessing on the signal z(k) to be processed to obtain a y(k) signal; Step 2, FFT transforming the y(k) signal through the time domain nonlinear preprocessing into the frequency domain to obtain the frequency domain signal Y(k); Step 3, performing hard decision processing in the frequency domain to obtain the signal Z'(k); Step 4, performing IFFT transformation on the signal Z'(k) after the hard decision to obtain the time domain signal y'(k); Step 5, obtain signal y(k) with step 1 and subtract step 4 to obtain y'(k) signal and obtain n'(k); Step 6, using the signal n'(k) to calculate the negative impulse noise estimate i'(k); Step 7. Subtract the negative impulse noise estimation i'(k) signal obtained in step 6 from y(k) obtained in step 1, and obtain the signal z'(k) after joint suppression of impulse noise. 2. The joint suppression method of impulse noise in the multi-carrier communication system according to claim 1, characterized in that, the time-domain nonlinear preprocessing in the step 1 adopts the method of zeroing. 3. by the joint suppression method of impulse noise in the described multi-carrier communication system of claim 1, it is characterized in that, the concrete steps of described step 6 are: First, calculate the average power of the signal n'(k) as S, $\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; =</span>\sqrt{\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>{<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; ;</span>$ Second, the negative impulse noise is estimated as: