A method for tolerance aided interference degradation for optical OFDM in power limited systems

Abstract

The application provides a tolerance aided interference degradation method for optical OFDM in a power limited system, which is based on CEO-OFDM, and proposes a TAID algorithm, which reduces the clipping interference and additive noise in the CEO-OFDM system by combining less signals from multiple frames when reconstructing data, sets the optimal decision tolerance at the receiver, and judges whether the signals in the clipping part need to be combined, so that the interference caused by the limited transmitting power is effectively reduced, and the noise level at the receiver is significantly reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of signal processing and communication signal detection, and particularly relates to a margin aided interference mitigation method for optical OFDM in a power limited system. BACKGROUND

[0002] Wireless optical communication (OWC) has received extensive attention from both industry and academia due to its more advantages compared with radio frequency (RF) communication systems. By using light emitting diodes (LEDs) as transmitters, OWC is not subject to RF interference, can provide higher security and data rates, and has lower power consumption than RF systems.

[0003] Orthogonal frequency division multiplexing (OFDM) is increasingly widely used in optical communication systems due to its resistance to inter-symbol interference (ISI) and high spectral efficiency. In OWC systems, intensity modulation and direct detection (IM / DD) is used, and the traditional OFDM scheme in radio frequency cannot be directly applied to OWC systems. Direct current biased optical OFDM (DCO-OFDM) is one of the most commonly used optical OFDM schemes.

[0004] For DCO-OFDM, a constant bias needs to be added to the bipolar OFDM signal to make the transmitted signal positive. Asymmetrically clipped optical OFDM (ACO-OFDM) does not need to add a direct current bias. Only odd subcarriers are used for modulation, but ACO-OFDM loses half of the bandwidth efficiency compared with DCO-OFDM. Flip OFDM, also known as U-OFDM, transmits positive and negative signals in independent adjacent frames, and the spectral efficiency is only half of that of DCO-OFDM.

[0005] Due to the peak power limitation of the light source, clipping distortion needs to be considered, which will significantly reduce the system performance. At present, a clipping enhanced optical OFDM (CEO-OFDM) for LED-based OWC systems has been proposed in the prior art, in which the clipped part of the signal is transmitted using an additional signal frame. Under the use of this transmission mode, the non-linear distortion caused by the peak power limitation is significantly reduced, and a lower bit error rate can be obtained than other OFDM schemes. However, CEO-OFDM introduces more noise when reconstructing the received signal, and more bandwidth is required when transmitting additional signal frames, which leads to the introduction of more noise at the receiver. SUMMARY

[0006] In view of the above problems, the application provides a tolerance aided interference degradation method for optical OFDM in a power limited system, which provides a tolerance aided interference degradation (TAID) algorithm, reduces the clipping interference and additive noise in a CEO-OFDM system by combining less signals from multiple frames when reconstructing data, and sets a decision tolerance for judging whether the received signal needs to include the signal in the clipping part.

[0007] The technical solution for achieving the object of the application is:

[0008] A tolerance aided interference degradation method for optical OFDM in a power limited system, characterized in that the method comprises the following steps:

[0009] Step 1: constructing a transmitter of CEO-OFDM, bipolar signal x s [m] of the transmitter is represented as:

[0010]

[0011] Wherein, P max is a radiation power peak value; N is the number of subcarriers; is a positive signal; is the negative signal taken inversely; is a clipping signal; represents the nonlinear response of the light source with peak power constraint;

[0012] And

[0013] Wherein, x is signal power;

[0014] Step 2: constructing a receiver of CEO-OFDM, the initial received mth discrete time signal of the receiver is:

[0015]

[0016] Wherein, ρ represents the photoelectric power conversion ratio of a photodetector; n y [m] is a Gaussian white noise with a mean of zero; h[m] is a system impulse response, and the operation "*" represents a discrete time convolution;

[0017] Step 3: setting an optimal decision tolerance at the receiver based on the TAID algorithm, judging whether the clipping signal is superimposed on the clipped signal through the optimal decision tolerance, and finally obtaining the reconstructed received signal.

[0018] Compared with the prior art, the application has the following beneficial effects:

[0019] The present application is based on the existing CEO-OFDM, and designs a decision threshold of a receiving end for reducing noise and interference. Since the CEO-OFDM uses an additional frame to transmit a clipping part of an OFDM signal to reduce clipping distortion, the reconstruction of the signal from all frames increases the noise level. The present application uses the decision threshold through the TAID algorithm, and determines whether the signal in the clipping part is necessary based on the signal amplitude in the first two frames using the CEO-OFDM. Then, the decision threshold is optimized, so that the clipping distortion and the additional noise contained are optimized, and the received signal-to-noise ratio is maximized. Finally, the experimental results prove that the TAID algorithm has better BER performance compared with the DCO and the CEO-OFDM. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 A schematic diagram of a CEO-OFDM received signal is shown in FIG. 1.

[0021] Figure 2 A receiving block diagram of the TAID algorithm is shown in FIG. 2.

[0022] Figure 3 A comparison of the bit error rate under different modulation indexes is shown in FIG. 3.

[0023] Figure 4 A comparison of the bit error rate of the TAID algorithm under different clipping detection thresholds is shown in FIG. 4. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solutions of the present application, the technical solutions of the present application are further described below in combination with the drawings and examples.

[0025] I. Based on the principle of CEO-OFDM

[0026] The clipping enhanced optical OFDM (CEO-OFDM) is similar to the DCO, the ACO and the U-OFDM, and needs to construct a Hermitian symmetric matrix to ensure that the transmitted signal is a real value. However, the bipolar signal generated through the inverse fast Fourier transform (IFFT) may be distorted due to the peak transmit power limit and the high peak-to-average power ratio (PAPR) of the signal. Therefore, in the CEO-OFDM system, 3 frames are used to transmit each OFDM symbol.

[0027] 1. Transmission signal

[0028] For simplicity, the transmission signal within an OFDM symbol time is analyzed here. In a multicarrier modulation system, the OFDM symbol time T = NT b , where N is the number of subcarriers, T b is the duration of a single symbol, and the transmission signal form is as shown in formula (3).

[0029] To achieve high spectral efficiency, M-QAM modulation is used for each OFDM symbol. After M-QAM modulation, the data of the i-th subcarrier is X i To ensure the OFDM signal is real, the data of the i-th subcarrier X i and the data of the N-1-i-th subcarrier X N-1-i must satisfy the conjugate symmetry, where N is the number of subcarriers.

[0030] After IFFT operation and serial-to-parallel conversion at N points, the m-th sample of each OFDM symbol is represented as:

[0031]

[0032]

[0033] where c is the modulation index of the control signal amplitude ratio, x s [m] represents the bipolar signal, k is the k-th time domain signal, and δ is the discrete impulse signal.

[0034] In the present system, LEDs are used as emitters, driven by a forward current that is easily modulated. The output optical power has a nonlinear relationship with the driving current, and the maximum optical power is limited by current saturation, which can cause severe clipping distortion. Below the maximum optical power, the relationship between the input and output of the LED can be linearized using an assumed pre-distorter. The optical power is modeled as being directly proportional to the input current in the range of 0 to P max The c is optimized by prior art to balance the signal power and the clipping distortion caused by the peak radiant power limit of the light source.

[0035] In the CEO-OFDM system, the bipolar signal x s [m] is transmitted in three parts, the first part is the positive signal, the second part is the negative signal taken as the opposite, and the third part is the clipping signal, which requires the use of 3 frames to send during transmission, so the transmission signal can be represented as:

[0036]

[0037]

[0038] where, represents the nonlinear response of the light source with a peak power constraint; P max is the radiant power limit.

[0039] 2. Receiving signal and noise

[0040] At the receiver, the mthdiscrete-time signal initially received can be modeled as:

[0041]

[0042] where ρ represents the photoelectric power conversion ratio of a photodetector (PD); n y [m] is additive noise, including thermal noise and shot noise; n y [m] is a Gaussian white noise with a mean of zero; h[m] is a system impulse response, and the operation "*" represents a discrete-time convolution;

[0043] and n y [m] has a variance calculation formula of R s is a transmission QAM symbol rate, and N0is a noise power spectral density;

[0044] If the amplitude of the extra frame is greater than the constrained power, the signal transmitted in the extra frame will still be amplitude-limited, and thus the transmitted data will be distorted. This distortion can be modeled as a Gaussian-distributed random variable with a mean of zero, and the variance of this distortion is a function of the modulation index c:

[0045]

[0046] where σ d is the variance of , is a complementary error function.

[0047] Second, the TAID algorithm

[0048] Although the CEO-OFDM system reduces the clipping distortion, it uses more bandwidth and introduces more interference and noise in the signal reconstruction process at the receiver compared with the DCO-OFDM. Therefore, in order to reduce the interference caused by the limited transmission power, the TAID algorithm sets an optimal decision threshold at the receiver to determine whether the clipped signal needs to be superimposed on the clipped signal.

[0049] 1. Theoretical analysis

[0050] As shown in Figure 1 , the steps of the TAID algorithm are as follows:

[0051] Step 1: Set the optimal decision threshold Δ;

[0052] Step 2: Determine whether the first frame and the second frame of the received signal are greater than P max - Δ, if yes, set the first frame and the second frame signal power to P maxand the third frame signal needs to be added to the first and second frames; if P max - Δ, then no power peak distortion is considered and no signal in the third frame is needed.

[0053] According to the TAID algorithm, the mth sample of the reconstructed signal is expressed as:

[0054] r[m] = r s [m] + r c [m], m = 0, 1,..., N - 1 (7)

[0055] where r s [m] is the clipped signal, i.e., the first and second frames, and r c [m] is the clipped signal, i.e., the third frame.

[0056] where:

[0057] r s [m] = (φ Δ (y[m]) - φ Δ (y[m + N]) U N [m] (8)

[0058] r c [m] = G(r s [m]) sign(y[m] - y[m + N]) · y[m + 2N] U N [m] (9)

[0059] where U N [m] = u[m] - u[m - N] is a rectangular sequence, u[m] is a unit step sequence, φ Δ (·) is used to limit the power of the received clipped signal (the first and second frames) to -P max + Δ ≤ x ≤ P max + Δ, G(·) is used to limit the power of the received clipped signal (the third frame) to -P max ≤ x ≤ P max , and the definitions of φ Δ (·) and G(·) are as follows:

[0060]

[0061]

[0062] where x is the signal power and Δ is the optimal decision margin.

[0063] The demodulated M-QAM signal on the ith subcarrier is modeled as:

[0064]

[0065] Where α(c) and |H i |、n i Let represent the clipping factor, the amplitude of the i-th subcarrier, and the additive noise, respectively. According to Bussgang's theorem, the clipping factor is calculated as follows:

[0066]

[0067] By using decision tolerance, the third frame signal can be made... The additive noise in the receiver is reduced.

[0068] During transmission, the three frames of signal pass through the channel separately, and each frame is subject to additive noise interference. Therefore, the noise in the three parts is modeled as a combination of multiple parts. According to the Central Limit Theorem (CLT), the effective noise n on the i-th subcarrier... i It can be viewed as a Gaussian distributed random variable with zero mean. This leads to the addition of [something] to the recovery signal. The formula for calculating the noise variance is as follows:

[0069]

[0070] The formula for calculating the effective noise component in formula (14) is as follows:

[0071]

[0072] Where q(r) and g(n) represent y[m] and n as shown in equation (5), respectively. y The probability density function of [m].

[0073] For a given modulation index and transmit symbol rate, the received signal-to-noise ratio (SNR) of the i-th subcarrier can be calculated as follows:

[0074]

[0075] Where H is the Fourier transform of h, and H[i] represents the frequency response of the i-th subcarrier;

[0076] Given the SNR, the bit error rate (BER) of the i-th subcarrier can be approximated as:

[0077]

[0078] 2. Specific Implementation

[0079] Due to the characteristics of LEDs, the light source limits the emitted signal power to a peak value P. max The following, such as Figure 1 As shown, greater than P max The received signal is inevitably affected by additive noise, Pmax Some signals in frames 1 and 2 below may also be affected by additive noise. In the TAID algorithm, an optimal decision tolerance Δ is first set to determine whether the signals in frame 3 need to be combined. If the signals in frames 1 and 2 are greater than P... max Then set the signal power of the first and second frames to P. max The signal in frame 3 corresponds to the clipped portion. If the signal in frames 1 and 2 is less than P... max If so, it is assumed that there is no peak power distortion and the corresponding clipping portion in the third frame is not needed. The entire transmission process is as follows: Figure 2 As shown.

[0080] Example

[0081] To further illustrate the effectiveness of the TAID algorithm proposed in this invention, the system performance of the proposed TAID algorithm is simulated and analyzed, and the DCO, CEO and BER performance of the TAID algorithm proposed in this invention are compared.

[0082] First, to simplify the problem, we assume that the channel is an ideal channel (zero loss) and set the simulation parameters with reference to Table 1.

[0083] Table 1 Simulation Parameters

[0084]

[0085] Appendix Figure 3 The figure shows a comparison of the BER performance of the proposed TAID algorithm with the DCO and CEO algorithms, where the TAID algorithm has been selected with the optimal clipping detection tolerance. As can be seen from the figure, the BER performance of all algorithms decreases with increasing modulation index. Further increasing the signal amplitude leads to more clipping distortion. When the clipping distortion exceeds the additive noise, the BER performance becomes even worse. Because CEO-OFDM uses additional signal frames to store and transmit clipped information, its bit error rate performance is better than DCO-OFDM. Since the TAID algorithm uses a clipping detection tolerance at the receiver, its BER is two orders of magnitude higher than that of CEO-OFDM. Compared to CEO-OFDM, the TAID algorithm requires a smaller modulation index to achieve the minimum BER, indicating that the TAID algorithm has higher power efficiency than CEO-OFDM.

[0086] Because the Tolerance-Assisted Interference Degradation Algorithm (TAID algorithm) proposed in this invention uses an optimal tolerance value at the receiver to minimize interference caused by peak power constraints and additive noise, such as Figure 4 As shown, the BER performance of the TAID algorithm with different tolerance parameters was compared. Figure 4As can be seen, for a certain modulation index, there is an optimal threshold value that can be found to achieve the minimum BER. When the modulation index is small, the signal clipping interference is not strong enough to add to the additive noise, and increasing the threshold value does not improve the BER. Further increasing the threshold value introduces more interference and results in performance degradation.

[0087] The content not described in detail in the specification belongs to the prior art known to the person skilled in the art. Although the patent of the present application has been described in detail with reference to the foregoing embodiments, the person skilled in the art can still modify the technical solutions recorded in the foregoing embodiments or make equivalent replacement to part of the technical features, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for tolerance-aided interference mitigation for optical OFDM in power-limited systems, characterized in that, The method comprises the following steps: Step 1: Constructing the transmitter of CEO-OFDM, and obtaining its bipolar signal x s The transmission signal of [m] is: where P max is the peak value of the radiation power; N is the number of subcarriers; is the positive signal; is the negative signal negated; is the clipping signal; denotes the nonlinear response of the light source with peak power constraint; and Wherein, x is signal power; Step 2: Constructing a receiver of CEO-OFDM, and obtaining an initial received mth discrete-time signal of the receiver as where p represents the photoelectric power conversion ratio of the photodetector; n y [m] is a Gaussian white noise with zero mean; h[m] is the system impulse response, and the operation "*" represents a discrete-time convolution; Step 3: Setting an optimal decision threshold at the receiver based on the TAID algorithm, judging whether to superimpose the clipped signal on the clipped signal through the optimal decision threshold, and finally obtaining a reconstructed received signal; The specific operation steps of step 3 are: Step 31: Setting an optimal decision threshold Δ; Step 32: Determine if the first frame of the received signal and the second frame of the received signal is greater than P max -△, if so, proceed to step 33, otherwise, assume no power peak distortion; Step 33: set the power of the first frame signal and the second frame signal to P max and superimpose the clipping signal onto the first frame signal and the second frame signal; Step 34: After the reconstruction processing of the received signal in step 33, the mth sample in a symbol is represented as r[m] = r s [m] + r c [m], m = 0, 1,..., N - 1 (7) where r s [m] is the clipped signal; r c [m] is the clipped signal; r s [m] = (φ △ (y[m])-φ △ (y[m+N])) U N [m] (8) r c [m] = G(r s [m]) sign(y[m] - y[m+N]) · y[m+2N] U N [m] (9) where U N [m] = u[m] - u[m - N] denotes a rectangular sequence, u[m] denotes a unit step sequence, Step 35: Transform the subcarriers of the reconstructed received signal with M-QAM, X i is the transform result: where a(c), |H i |, n i denote the clipping coefficient, the amplitude of the i-th subcarrier and the additive noise, respectively. The received signal-to-noise ratio SNR of each subcarrier is: Given SNR, the bit error rate BER of the subcarrier is calculated through formula (17):

2. A margin assisted interference mitigation method for optical OFDM in power limited systems as claimed in claim 1 characterized by: The X i The noise variance on the X The effective noise component in formula (14) is: where q(r) and g(n) represent the probability density functions of y[m] and n y [m] respectively.

3. A margin assisted interference mitigation method for optical OFDM in power limited systems as claimed in claim 2, characterized by: The calculation formula of the clipping coefficient α(c) is: where c is a modulation index of the control signal amplitude ratio; σ d is X i variance.

Images