This article uses the parabolic model in to estimate the parameters of distortion based on the frequency domain pilot signal in the OFDM system, thereby recovering the distorted signal and performing distortion cancellation. The simulation results show that compared with traditional methods, the new method proposed in this paper has a performance gain of about 2dB under high signal-to-noise ratio.

1 System model

The transmitter of the OFDM system is shown in Figure 1. The transmitter first maps binary sources to complex points on a fixed constellation diagram and converts them into parallel data streams. The number of parallel data for each OFDM symbol is determined by the number of subcarriers in the system. Then insert a pilot signal with a predetermined position and size in the middle, as the specified pilot position. The information sent by these pilot signals is known to the receiver and can be used to estimate the impact of the external environment on the transmitted signal, such as time-varying channel effects. This article uses them to estimate distorted signals. Transform the data stream into a time-domain signal through IFFT operation, and finally convert it into a serial data stream and send it out as an analog signal through a digital to analog converter and power amplifier, as shown in Figure 1.

For hard limiting systems, signal distortion can be modeled as a parabola with random parameters, as shown in Figure 2.

Methods for producing distortion cancellation

To cancel out distorted signals, we first use known information to estimate the distorted signal we obtain, and then have different effects on the distortion caused by the cancellation in the original signal:

The key is how to make good recovery for distorted signals in OFDM systems. Below, we will use the parabolic method to study the model and solve the problem.

Firstly, the unique frequency domain characteristics of distorted signals were analyzed. Based on the analysis of the parabolic model, we can make DFT transformations to obtain specific expressions of frequency domain estimation by analyzing the relevant distorted signals.

The impact of random variables, as we understand it, is mainly on amplitude, but the most significant impact is still on phase.

First, the duration of frequency domain estimation.

The method of calculating the minimum mean square error (MMSE) using the minimum mean method is used to estimate the frequency domain from the points of the sample, that is, to find multiple estimated values in the frequency domain and calculate the result.

Finally, through calculation, using the distorted signal estimated in the frequency domain that has already been recovered, we can cancel out the influence of distortion from the received frequency domain estimated signal and restore the original signal.

Finally, it should be noted that since the method used in the derivation is approximate, it is necessary to meet the required conditions as much as possible. When estimating in the frequency domain, the number of derivatives taken in the frequency domain should meet the conditions for estimating the frequency domain as much as possible to ensure the accuracy of the estimated data. We also tried to use subcarrier points with lower frequency estimates for estimation.

3. Simulation results and analysis

The simulation uses an OFDM system in 2048 subcarrier data, which is modulated using 16QAM. The original transmitted information is evenly distributed on the basic points of the constellation. The ratio of the threshold for limiting the amplitude of the amplifier to the average amplitude of the input signal is 4.5dB. In the region with a signal-to-noise ratio of 6dB and 14dB, we understand the symbol error rate (SER) curves of the algorithm proposed in this paper and the algorithms mentioned in the literature, as shown in Figure 3.

The three curves that appear in Figure 3 are as follows: DYSF118B 61430001-XG uses the algorithm proposed in this paper to cancel out the distortion and show the performance curve. The literature used the method shown in the paper to cancel out the distortion and show the performance curve of the system without causing distortion. It can be understood that the method described in the literature safely cancels out the nonlinearity of the distortion and has an impact on the performance of the system where the distortion is located. Compared to the distortion cancellation that was not used, it results in a gain of 1.5dB to 2dB performance for the system. Compared to the low signal-to-noise ratio method, the method shown in this article and the methods mentioned in the literature have higher similarities in performance. As the signal-to-noise ratio method is improved, the performance of the new method described in this article is significantly better than that of the methods mentioned in the literature. When the signal-to-noise ratio is 14dB, the method shown in this article has a better performance in the literature, with a performance of 2dB.

The main reason that affects the performance of the algorithm is actually the presence of noisy noise in the signal received from the pilot. In the literature explanation method shown, it provides a concise definition of distortion, which is an average of the entire signal contained in the frequency domain estimation results of a linear function, which must to some extent bring some errors to the eliminated noise. So in this method, each point is not sensitive to instantaneous noise, and the difference in performance gain under high signal-to-noise ratio and low signal-to-noise ratio is not significant. The method shown in this article represents the frequency domain estimation of parameters that depend on distortion, and has a relatively close relationship with accuracy and noise. So for high signal-to-noise ratios, the estimation of distortion should be more accurate, and there is a conclusion that the performance gain of its occurrence is significantly better than that of low signal-to-noise ratios.

4 Conclusion

OFDM systems have a high peak to average ratio. Due to the limited linear range of the power amplifier, DYSF118B 61430001-XG will cause serious signal distortion. This article studies the cancellation method for this distortion. By modeling the distortion as a parabolic model, this paper analyzes its frequency domain characteristics and derives two parameters of distortion: initial time and duration, which affect the phase and amplitude of the frequency domain signal, respectively. Based on this, this article uses the minimum mean square error criterion to estimate the distortion parameters in the frequency domain, ultimately recovering the distorted signal and performing distortion cancellation. The simulation results show that under low signal-to-noise ratio, the method described in this paper has approximately the same performance as the method in the literature, and as the signal-to-noise ratio increases, the new method in this paper has a performance gain of about 2dB compared to the method shown in the literature.