Flexible NOMA-based NOHO-OFDM scheme for visible light communication with iterative interference cancellation

Xuan Huang; Fang Yang; Fang Yang; Changyong Pan; Changyong Pan; Jian Song; Jian Song

doi:10.1364/OE.420848

1. Introduction

Visible light communication (VLC) has become an attractive research field in recent years due to its advantages of ubiquitous deployment, low cost, broad bandwidth, and high security [1]. Orthogonal frequency division multiplexing (OFDM) have gained much attention in the VLC system because of its high spectral efficiency [2]. However, the OFDM method can not be immediately used to the VLC systems because the optical signals are required to be non-negative and real in VLC [3]. The real values can be ensured by imposing Hermitian symmetry to the frequency-domain signals [4]. To ensure the non-negative characteristic, asymmetrically clipped optical OFDM (ACO-OFDM) [5], DC biased optical OFDM (DCO-OFDM) [6], and pulse-amplitude-modulated discrete multitone (PAM-DMT) [7] have been proposed. However, the ACO-OFDM scheme and the PAM-DMT scheme only modulate the odd subcarriers and the imaginary part of the subcarriers, respectively, so that the spectrum efficiency of ACO-OFDM and PAM-DMT is low [8]. In DCO-OFDM, the DC bias does not carry any effective information, so the optical power efficiency is low [9].

Thus, several hybrid methods have been investigated to enhance the conventional schemes [10]. Asymmetrically clipped DC biased optical OFDM (ADO-OFDM) transmits ACO-OFDM signals on the odd subcarriers and DCO-OFDM signals on the even subcarriers to overcome the drawbacks of low spectral efficiency of ACO-OFDM and low optical power efficiency of DCO-OFDM [11]. A hybrid ACO-OFDM (HACO-OFDM) scheme is achieved by modulating ACO-OFDM signals and PAM-DMT signals on the odd subcarriers and the imaginary part of the even subcarriers, respectively [12]. To improve the spectral efficiency, the layered ACO-OFDM (LACO-OFDM) is proposed by utilizing different layers of ACO-OFDM signals [13], while the enhanced unipolar OFDM (eU-OFDM) is proposed by utilizing multiple unipolar data streams [14].

In order to further improve the achievable data rate, non-orthogonal multiple access (NOMA) is considered to apply to the VLC system [15]. In NOMA, each user can utilize the whole or partial bandwidth in the whole or partial time [16]. Users adopt superposition coding at the transmitter and successive interference cancellation (SIC) at the receiver to distinguish users through the power domain [17]. In [18], a power allocation scheme based on the differential evolution is proposed to achieve a tradeoff between the sum data rate and the user fairness. Another way to improve the performance of the NOMA-based VLC system is optimizing the constellation of the superposition coding [19]. In [20], the superposition constellation is adjusted to decrease the bit error rate (BER), while the symmetric superposition coding is proposed to eliminate the error propagation in [21].

However, in current NOMA-based VLC systems, signals from multiple users are overlapped on all subcarriers, which will increase the detection noise and can not meet different requirements of various users flexibly. Moreover, in current NOMA-based VLC systems, in addition to the channel noise and the interference from other users, there will be clipping noise caused by the user’s own signals, which will increase the interference in the process of SIC, then reducing the system performance. Thus, in this paper, a novel modulation scheme with iteratively successive interference cancellation (ISIC) is proposed. Compared to conventional modulation scheme, the proposed scheme has the advantages of superior flexibility, high throughput, and satisfactory performance. Specifically, the main contributions of this paper are listed as follows.

• A flexible NOMA-OFDM-based modulation scheme called non-orthogonal hybrid optical OFDM (NOHO-OFDM) is proposed to improve the achievable data rate region. The numbers of subcarriers allocated to different modulation methods can be adaptively changed according to users’ demands, thus enhancing the flexibility of the system.
• By making full use of the characteristics of time-domain signals, an ISIC scheme is proposed to reduce the impact of interference and noise, thus improving the BER performance of the system.
• Both simulation and experiment results indicate that the BER performance of the NOHO-OFDM scheme can be significantly improved by the proposed ISIC method, while only one iteration is required. Meanwhile, simulation results show that the achievable data rate region of the proposed NOHO-OFDM scheme is greatly enlarged compared with that of the orthogonal counterpart.

The rest of this paper is organized as follows. Section 2 introduces the proposed NOHO-OFDM scheme. In Section 3, the transmitter and receiver structures of the NOHO-OFDM scheme are investigated, and the corresponding achievable data rate of the proposed scheme is analyzed. Experiment and simulation results are given in Section 4. Finally, the conclusions are drawn in Section 5.

2. NOHO-OFDM with subcarrier allocation

In this section, a novel NOMA-based modulation scheme called NOHO-OFDM is proposed, in which subcarriers can be flexibly and non-orthogonally allocated. The NOHO-OFDM scheme can be regarded as a hybrid combination of overlapped ACO-OFDM and DCO-OFDM signals. Compared with the ADO-OFDM scheme, the differences between the ADO-OFDM and NOHO-OFDM schemes are as follows.

First, in the proposed NOHO-OFDM scheme, the numbers of subcarriers allocated to ACO-OFDM and DCO-OFDM schemes are no longer fixed as $N/2$ and $N/2-2$, where $N$ represents the available number of subcarriers, but can be flexibly changed according to different demands of users. For example, it can be determined according to the numbers of users with different complexity and/or latency requirements. The subcarrier diagram of the proposed NOHO-OFDM scheme is illustrated in Fig. 1. In this instance, the numbers of subcarriers allocated to the ACO-OFDM and non-overlapped DCO-OFDM schemes, represented as $N_A$ and $N_D$, are $N/4$ and $3N/4-2$, respectively. As long as the ACO-OFDM signals are only modulated on the odd subcarriers, the clipping noise due to the asymmetrical clipping will only fall on even subcarriers, which will not interfere ACO-OFDM signals transmitted on odd subcarriers [22]. Thus, even if the subcarrier allocation is performed, the ACO-OFDM signals can be effectively estimated.

Second, inspired by NOMA, non-orthogonality is introduced into the proposed modulation scheme to further improve the spectral efficiency of the VLC system. In the NOHO-OFDM scheme, the DCO-OFDM signals can be transmitted not only on the subcarriers that do not convey ACO-OFDM signals, but also on the subcarriers that transmit ACO-OFDM signals. As shown in Fig. 1, for the ACO-OFDM signals, only portion of the odd subcarriers are utilized to transmit the ACO-OFDM signals, and the clipping noise due to the asymmetrical clipping will fall on all even subcarriers. For the DCO-OFDM signals, they can be divided into two portions, the non-overlapped and overlapped DCO-OFDM signals. The non-overlapped DCO-OFDM signals refer to the DCO-OFDM signals transmitted on the subcarriers which do not transmit the ACO-OFDM signals, as shown by the solid lines with blue squares, while the overlapped DCO-OFDM signals refer to the DCO-OFDM signals conveyed on some of the subcarriers that transmit the ACO-OFDM signals, as shown by the solid lines with yellow circles.

Fig. 1. The subcarrier structure of the proposed NOHO-OFDM scheme.

Download Full Size | PDF

In the instance shown in Fig. 1, the number of subcarriers assigned for the overlapped DCO-OFDM signals is $N/8$, which is denote as $N_P=N/8$. The specific number of subcarriers assigned to the ACO-OFDM, non-overlapped DCO-OFDM, and overlapped DCO-OFDM can be adaptively changed according to the user’s demands, e.g., the data rate to be transmitted, thus increasing the flexibility of the VLC system. Then, the time-domain ACO-OFDM and DCO-OFDM signals are generated independently, and the NOHO-OFDM signals can be obtained by adding them together.

For the combined signals, as aforementioned, some ACO-OFDM signals and DCO-OFDM signals will be transmitted on the same subcarriers. Therefore, in order to demodulate the overlapped DCO-OFDM and ACO-OFDM signals successfully at the receiver, different powers will be allocated to the overlapped DCO-OFDM and ACO-OFDM signals, so that the SIC method can be used for demodulation at the receiver more efficiently [23], which will be explained in detail later.

3. Transmitter and receiver structures

In this section, the corresponding transmitter and receiver structures of the flexible NOHO-OFDM scheme are investigated. After that, the achievable data rate region of the proposed scheme is analyzed.

3.1 Transmitter structure

The transmitter structure of the proposed NOHO-OFDM scheme is shown in Fig. 2. According to the user’s requirements, subcarrier allocation is carried out in advance. Denote the index sets of subcarriers allocated to ACO-OFDM and DCO-OFDM schemes as ${{\boldsymbol {\zeta }}_{aco}}$ and ${{\boldsymbol {\zeta }}_{dco}}$, respectively. ${{\boldsymbol {\zeta }}_{dco}}$ can be further divided into the non-overlapped DCO-OFDM set, ${{\boldsymbol {\zeta }}_{ndco}}$, and the overlapped DCO-OFDM set, ${{\boldsymbol {\zeta }}_{odco}}$. And to improve the effectiveness of the SIC scheme, power allocation is performed beforehand.

Then, the generated bit streams will perform the operations of constellation mapping, serial-to-parallel (S/P), and the Hermitian symmetry imposing in turn to obtain the ACO-OFDM, non-overlapped DCO-OFDM, and overlapped DCO-OFDM symbols to be transmitted, i.e., $X_k$, $Y_{k,1}$, and $Y_{k,2}$. After that, inverse fast Fourier transform (IFFT) and parallel-to-serial (P/S) are carried out to generate the time-domain signals $x_n$, $y_{n,1}$, and $y_{n,2}$, where $x_n$ represents the bipolar time-domain ACO-OFDM signals. And the bipolar time-domain DCO-OFDM signals can be obtained by

(1)$$y_n = y_{n,1}+y_{n,2}.$$

Fig. 2. The transmitter structure of the proposed NOHO-OFDM scheme.

Download Full Size | PDF

After that, in order to obtain non-negative ACO-OFDM signals, $x_n$ is asymmetrically clipped as

(2)$$x_{aco,n} = \left\{ \begin{array}{ll} 0, &x_{n} \:<\: 0\\ x_{n}, & x_{n} \ge 0 \end{array} \right.,$$

while $y_n$ will go through the operations of adding DC component and zero clipping, which can be given as

(3)$$y_{dco,n} = \left\{ \begin{array}{ll} 0, &y_{n} + {B_{\textrm{DC}}} \:< \:0\\ y_{n} + {B_{\textrm{DC}}}, &y_{n} + {B_{\textrm{DC}}} \ge 0 \end{array} \right.,$$

where ${B_{\textrm{DC}}}$ represents the DC bias to be added. Then, the unipolar ACO-OFDM signals $x_{aco,n}$ and unipolar DCO-OFDM signals $y_{dco,n}$ are added together to obtain the hybrid combined NOHO-OFDM signals $z_{noho,n}$. Finally, the digital-to-analog (D/A) conversion is performed, and the signals are modulated onto the illumination intensity of the LEDs for transmission.

3.2 Receiver structure

Because the ACO-OFDM and DCO-OFDM signals are superimposed on portion of subcarriers, the SIC scheme is adopted in this paper. Specifically, the ACO-OFDM signals are demodulated according to the demodulation method of the conventional ACO-OFDM first. In this process, the overlapped DCO-OFDM signals should be treated as noise, thus, the power allocated to the overlapped DCO-OFDM signals should be less than that of the ACO-OFDM signals to demodulate signals more accurately. After that, the clipping distortion of the ACO-OFDM signals can be estimated and subtracted. Then, the DCO-OFDM signals can be recovered.

However, the overlapped DCO-OFDM signals will interfere the demodulation of the ACO-OFDM signals. In turn, the recovery accuracy of the ACO-OFDM signals will affect the demodulation of the DCO-OFDM signals, so the BER performance of the NOHO-OFDM scheme is worse than that of the orthogonal counterpart. In view of this, by making full use of the inherent characteristic of NOHO-OFDM signals in the time domain, a more accurate receiver with the ISIC scheme is proposed in this subsection.

Specifically, the inherent characteristic to be utilized is the antisymmetry characteristic of the unclipped time-domain ACO-OFDM signals [24], which can be formulated as

(4)$${x_n} ={-} {x_{n + N/2}},~~~{0 \le n \:<\: N/2}.$$

Thus, after asymmetrical clipping, one of $x_{aco,n}$ and $x_{aco,n + N/2}$ must be zero. Based on this, pairwise clipping is adopted at the receiver to decrease the influence of noise and interference. Specifically, for $n=0,1,\ldots ,{N}/{2}-1$, the updated time-domain ACO-OFDM signals can be estimated as

(5)$$\overline x _{aco,n} = \left\{ \begin{array}{ll} 0, &\widehat x_{aco,n} \:<\: \widehat x_{aco,n + N/2}\\ \widehat x_{aco,n}, &\widehat x_{aco,n} \ge \widehat x_{aco,n + N/2} \end{array} \right.,$$

and

(6)$$\overline x _{aco,n + N/2} = \left\{ \begin{array}{ll} \widehat x_{aco,n + N/2}, & \widehat x_{aco,n} \:<\: \widehat x_{aco,n + N/2}\\ 0, &\widehat x_{aco,n} \ge \widehat x_{aco,n + N/2} \end{array} \right..$$

For indoor VLC systems, the channel conforms to be the Lambertian model, in which the optical power received from reflected signals is always small and negligible [25]. Thus, it is a common strategy to concentrate on the line-of-sight (LoS) link, which means the influence of ISI can be ignored. In addition, to resist the dispersive channel of broadband system, preamble is usually used for the synchronization and channel estimation. Based on the results of channel estimation, the ISI can be eliminated in advance by a one-tap equalizer. Therefore, the minimum of the two signal samples in (5) and (6) could be set to zero to reduce the impact of noise.

The receiver structure of the NOHO-OFDM scheme with the proposed ISIC scheme is shown in Fig. 3. Take the $i$-th iteration as an example, the operation process of the receiver can be divided into the following five steps.

Fig. 3. The receiver structure of the proposed NOHO-OFDM scheme with ISIC.

Download Full Size | PDF

Step 1: Preparation for the iteration.

After passing through the analog-to-digital (A/D) converter, the received signals $r_{noho,n}$ will go through the operations of S/P and $N$-point FFT to generate the frequency-domain signals $R_k$, which are regarded as the input of the iterative process. For the convenience of understanding, $R_k$ is marked as $\overline X_{aco,k}^{(0)}$.

Step 2: Regenerate the time-domain DCO-OFDM signal.

First, the frequency-domain ACO-OFDM signals are estimated by $\overline X_{aco,k}^{(i-1)}$ according to the ML criterion, which can be given as

(7)$$\widetilde X_k^{(i)} = \mathop{\textrm{argmin}}_{X \in {\Omega _X}} ||X - 2\overline X_{aco,k}^{(i - 1)}|{|^2},~0<k<\frac{N}{2}~\textrm{and}~k \in {{\boldsymbol{\zeta}}_{aco}},$$

where ${\Omega _X}$ represents the constellation set of the ACO-OFDM scheme. Then, the operations of Hermitian imposing, $N$-point IFFT, and asymmetrical clipping are carried out in turn to generate the estimations of the time-domain ACO-OFDM signals, i.e., ${\widetilde x_{aco,n}^{(i)}}$. After that, the time-domain DCO-OFDM signals are regenerated as

(8)$${\widetilde y_{dco,n}^{(i)}} = {r_{noho,n}} - {\widetilde x_{aco,n}^{(i)}}.$$

Step 3: Calculate the estimation of frequency-domain DCO-OFDM signals.

The frequency-domain DCO-OFDM signals ${\widetilde Y_{dco,k}^{(i)}}$ are obtained after the operation of $N$-point FFT based on $\widetilde y_{dco,n}^{(i)}$. After that, the estimations of the non-overlapped and overlapped DCO-OFDM signals in the frequency domain are regenerated as

(9)$${\widetilde Y_{k,1}^{(i)}} = \mathop {{\mathop{\textrm{argmin}}\nolimits} }_{Y \in {\Omega}_Y } || {Y - {\widetilde Y_{dco,k}^{(i)}}} |{|^2},~0<k<\frac{N}{2}~\textrm{and}~k \in {{\boldsymbol{\zeta}}_{ndco}},$$

and

(10)$${\widetilde Y_{k,2}^{(i)}} = \mathop {{\mathop{\textrm{argmin}}\nolimits} }_{Y \in {\Omega}_Y } || {Y - {\widetilde Y_{dco,k}^{(i)}}} |{|^2},~0<k<\frac{N}{2}~\textrm{and}~k \in {{\boldsymbol{\zeta}}_{odco}},$$

where ${\Omega _Y}$ represents the constellation set of the DCO-OFDM scheme.

Step 4: Regenerate the time-domain ACO-OFDM signals.

First, based on ${\widetilde Y_{k,1}^{(i)}}$ and ${\widetilde Y_{k,2}^{(i)}}$ and the allocated power, the time-domain DCO-OFDM signal $\widehat y_{dco,n}^{(i)}$ can be obtained after the processes of Hermitian imposing, IFFT, and adding DC bias. Thus, the bipolar time-domain ACO-OFDM can be calculated as

(11)$$\overline{\overline x} _{aco,n}^{(i)} = {r_{noho,n}} - \widehat y_{dco,n}^{(i)}.$$

After that, asymmetrical clipping is performed based on $\overline {\overline x} _{aco,n}^{(i)}$ to regenerate the time-domain ACO-OFDM signals ${\widehat x_{aco,n}^{(i)}}$.

Step 5: Update the frequency-domain ACO-OFDM signals by the pairwise clipping.

Pairwise clipping is performed to obtain the updated time-domain ACO-OFDM signals $\overline x _{aco,n}^{(i)}$. Then, the frequency-domain ACO-OFDM signals $\overline X_{aco,k}^{(i)}$ can be regenerated after carrying out $N$-point FFT.

So far, the $i$-th iteration of the proposed ISIC scheme is finished. Using $\overline X_{aco,k}^{(i)}$ as the input of the $(i+1)$-th iteration, repeat Step 2 to Step 5 to obtain the updated ACO-OFDM and DCO-OFDM estimations until reaching the maximum iteration number $I$.

The computational complexity of both the conventional SIC and proposed ISIC is analyzed. For the conventional SIC scheme, two $N$-point FFT and one $N$-point IFFT should be performed, thus, the computational complexity can be calculated as $3\mathcal {O}(Nlog_{2}N)$. For the proposed ISIC scheme, two $N$-point FFT and two $N$-point IFFT are required for each iteration, thus, the computational complexity is $4I\mathcal {O}(Nlog_{2}N)$. Thus, it can be found that the computational complexity and the latency of the proposed ISIC scheme are both about $4I/3$ times of those of the conventional one.

3.3 Analysis for the achievable data rate

Assume that the total electrical power assigned to the NOHO-OFDM scheme is $P$, while $\alpha$ and $\beta$ are the power allocation factors for the ACO-OFDM signals and non-overlapped DCO-OFDM signals, respectively. Thus, the power of overlapped DCO-OFDM signals can be calculated as $\left ( {1{ - }\alpha { - }\beta } \right )P$.

Thus, the achievable data rates of the ACO-OFDM and DCO-OFDM portions in the NOHO-OFDM scheme can be formulated as

(12)$${C_A} = \frac{{\left( {{N_A} - N_P} \right)B}}{{2N}}{\log _2}\left( {1 + \frac{{\alpha P \cdot \frac{{{N_A} - N_P }}{{{N_A}}}}}{{{G_0}B \cdot \frac{{{N_A} - N_P }}{N}}}} \right)\\ + \frac{{N_P B}}{{2N}}{\log _2}\left( {1 + \frac{{\alpha P \cdot \frac{{N_P }}{{{N_A}}}}}{{\left( {1 - \alpha - \beta } \right)P + {G_0}B \cdot \frac{{N_P }}{N}}}} \right),$$

and

(13)$$C_{D}=\frac{N_{D} B}{2 N} \log _{2}\left(1+\frac{\beta P}{\varepsilon_{1} \cdot \alpha P+ G_{0} B\cdot\frac{N_{D}}{N} }\right)\\ +\frac{N_P B}{2 N} \log _{2}\left(1+\frac{(1-\alpha-\beta) P}{\varepsilon_{2} \cdot \alpha P \cdot \frac{N_P}{N_{A}}+ G_{0} B\cdot\frac{N_P}{N}}\right),$$

respectively, where $B$ represents the available transmission bandwidth, and $G_0$ denotes the power spectral density of the additive white Gaussian noise. $\varepsilon _{1}$ represents the error factor of clipping noise due to the estimation error of the ACO-OFDM signals. And $\varepsilon _{2}$ represents the interference factor caused by the estimation error of the ACO-OFDM signals to the overlapped DCO-OFDM signals. In order to simplify the analysis, the following simulation in this paper will make $\varepsilon _{1}=\varepsilon _{2}=\varepsilon$.

4. Experiments and simulations

In this section, experiments and simulations are performed to evaluate the performance of the proposed NOHO-OFDM with ISIC scheme.

Figure 4 shows the experimental setup of the VLC system. Original signals are generated by an arbitrary waveform generator (AWG, Rigol DG5352). Signals are added with a DC-bias voltage by a Tektronix Bias-Tee to ensure working in the linear range of the LED. After that, the combined signals are transmitted from LEDs and captured by a PD (Thorlabs PDA10A2) after a free-space transmission of 0.6 m. Then, the received signals are recorded by a digital serial analyzer (Tektronix DSA 71254C) for off-line processing. In this experiment, the size of FFT is 1024, QPSK is adopted, and the average output power is set as a constant for various scenarios.

Fig. 4. Experimental setup of the VLC system.

Download Full Size | PDF

Figure 5 shows the BER performance of ACO-OFDM and DCO-OFDM portions in the NOHO-OFDM scheme with $N_P = 256$. The horizontal axis represents data rates transmitted by the NOHO-OFDM scheme. Because the average output power is set as a constant, the BER will increase with the increase of data rate. The BER performance of both ACO-OFDM and DCO-OFDM portions can be significantly improved after adopting the proposed ISIC scheme, and the ISIC scheme can be basically converged after one iteration. When the data rate is low, the ISIC scheme can improve the BER performance of ACO-OFDM more effectively due to the high signal-to-noise ratio (SNR). Meanwhile, the BER performance of DCO-OFDM can be greatly improved because of the more accurate estimations of ACO-OFDM signals.

Fig. 5. Experimental results of data rate versus BER performance for ACO-OFDM and DCO-OFDM portions in the NOHO-OFDM scheme.

Download Full Size | PDF

Figure 6 shows the experimental results of data rate versus BER for the orthogonal scheme and the proposed NOHO-OFDM scheme. For any data rate, when the conventional SIC scheme is utilized, the BER performance of the NOHO-OFDM is worse than that of the orthogonal counterpart due to the increased interference caused by overlapped DCO-OFDM signals. However, after utilizing the proposed ISIC scheme, the BER of the NOHO-OFDM scheme with a data rate lower than 46.8 Mb/s can be significantly improved and is less than that of the orthogonal scheme. It is because the data rate in this experiment is adjusted by the occupied bandwidth, lower bandwidth will result in lower noise, and the NOMA scheme in high SNR scenario will bring more performance gains in terms of the BER.

Fig. 6. Experimental results of data rate versus BER for the orthogonal scheme and the proposed NOHO-OFDM scheme with $N_P = 256$.

Download Full Size | PDF

The achievable data rate regions of the orthogonal scheme and the proposed NOHO-OFDM scheme with different $N_P$ is figured in Fig. 7. The Lambertian model is used as the VLC channel mode for simulation. With the increase of $N_P$, the achievable data rate region is enlarged. Specifically, the maximum achievable data rates of ACO-OFDM with different $N_P$ are the same. It is because all the power is allocated to the ACO-OFDM scheme, which means that both orthogonal and non-orthogonal schemes are degenerated into ACO-OFDM schemes. In addition, with the increase of $N_P$, the maximum achievable data rate of DCO-OFDM increases due to the increase of the available bandwidth for the DCO-OFDM scheme. It can be calculated that when $N_P =$ 128, 256, and 512, compared with the orthogonal scheme, the achievable data rate regions of the proposed NOHO-OFDM scheme are expanded by 12.9$\%$, 26.2$\%$, and 31.5$\%$, respectively.

In addition, the achievable data rate regions of the orthogonal scheme and the proposed NOHO-OFDM scheme with different $\varepsilon$ when $N_P$ is 256 are shown in Fig. 8. The achievable data rate region can be further enlarged with the decrease of the residual interference factor $\varepsilon$. When $\varepsilon = 1\%$ and $\varepsilon = 1$‰, the achievable data rate regions of the proposed NOHO-OFDM scheme are expanded by 21.7$\%$ and 30.2$\%$ compared with the orthogonal scheme, respectively.

Fig. 7. Comparison of the achievable data rate regions between the orthogonal scheme and the proposed NOHO-OFDM scheme with different $N_P$.

Download Full Size | PDF

Fig. 8. Comparison of the achievable data rate regions between the orthogonal scheme and the proposed NOHO-OFDM scheme with different $\varepsilon$.

Download Full Size | PDF

The BER performance of the ACO-OFDM portion in the NOHO-OFDM scheme with the conventional and iterative receivers is shown in Fig. 9. No matter what $N_P$ is, the BER performance of the ACO-OFDM portion can be significantly improved after utilizing the proposed ISIC scheme because of the adopted pairewise clipping, which can reduce the influence of noise and interference. It can be calculated that at the BER of $10^{-3}$, when $N_P =$ 128, 256, and 512, the proposed ISIC scheme can achieve 1.3 dB, 1.1 dB, and 0.8 dB gains compared with the conventional SIC scheme, respectively. In addition, for any $N_P$, the BER performance of the system is almost the same at $I = 1$ and $I = 2$, which means the proposed ISIC scheme can be basically converged after only one iteration.

Fig. 9. The BER performance of the ACO-OFDM portion in the NOHO-OFDM scheme with conventional SIC scheme and proposed ISIC scheme.

Download Full Size | PDF

The BER performance of the DCO-OFDM portion in the NOHO-OFDM scheme with the conventional and iterative receivers is shown in Fig. 10. Similar to the ACO-OFDM portion, for any $N_P$, the BER performance of the DCO-OFDM portion is improved after adopting the proposed ISIC scheme due to the more accurate estimations of ACO-OFDM signals. And at the BER of $10^{-3}$, when $N_P =$ 128, 256, and 512, the proposed ISIC scheme can achieve 0.5 dB, 0.5 dB, and 0.4 dB gains compared with the conventional SIC scheme, respectively.

Fig. 10. The BER performance of the DCO-OFDM portion in the NOHO-OFDM scheme with conventional SIC scheme and proposed ISIC scheme.

Download Full Size | PDF

The BER performance for the orthogonal scheme and the NOHO-OFDM scheme with the conventional and iterative receivers is shown in Fig. 11. The BER performance improvement of ACO-OFDM and DCO-OFDM portions will boost the performance of NOHO-OFDM. No matter what $N_P$ is, the BER performance of the orthogonal scheme is superior than that of the NOHO-OFDM scheme when the conventional SIC is adopted. With the iterative receiver, for $N_P = 512$ and $N_P = 256$, although the orthogonal scheme still outperforms the non-orthogonal one, the performance difference between these two schemes is greatly reduced when the SNR is high. And for $N_P = 128$, the BERs of the conventional and the proposed NOHO-OFDM schemes are almost the same when SNR $= 17$ dB. Furthermore, for any $N_P$, the proposed ISIC scheme can be basically converged after only one iteration.

Fig. 11. The BER performance of the orthogonal scheme and the proposed NOHO-OFDM scheme.

Download Full Size | PDF

5. Conclusions

In this paper, the NOHO-OFDM scheme is proposed to further improve the achievable data rate region with flexible subcarrier allocation. And an ISIC scheme is investigated to reduce the impact of noise and interference. Then, the achievable data rate of the proposed scheme is analyzed and formulated. Simulation results show the achievable data rate of the NOHO-OFDM scheme is much larger than that of the orthogonal counterpart. Experiment results show that the proposed NOHO-OFDM with ISIC scheme is superior than the conventional scheme in terms of the BER performance and data rate. Thus, it is seen that the proposed scheme can meet the requirements of superior flexibility, high throughput, and satisfactory performance, which can be served as a potential candidate for the signal modulation and detection in the VLC systems.

Funding

National Natural Science Foundation of China (61871255); National Key Research and Development Program of China (2017YFE0113300); Fok Ying Tung Education Foundation.

Disclosures

The authors declare no conflicts of interest.

References

1. T. Wang, F. Yang, J. Song, and Z. Han, “Dimming techniques of visible light communications for human-centric illumination networks: State-of-the-art, challenges, and trends,” IEEE Wireless Commun. 27(4), 88–95 (2020). [CrossRef]

2. X. Huang, F. Yang, X. Liu, H. Zhang, J. Ye, and J. Song, “Subcarrier and power allocations for dimmable enhanced ADO-OFDM with iterative interference cancellation,” IEEE Access 7, 28422–28435 (2019). [CrossRef]

3. A. Weiss, A. Yeredor, and M. Shtaif, “Iterative symbol recovery for power-efficient DC-biased optical OFDM systems,” J. Lightwave Technol. 34(9), 2331–2338 (2016). [CrossRef]

4. S. D. Dissanayake, K. Panta, and J. Armstrong, “A novel technique to simultaneously transmit ACO-OFDM and DCO-OFDM in IM/DD systems,” in 2011 IEEE GLOBECOM Workshops (GC Wkshps), (2011), pp. 782–786.

5. L. Chen, B. Krongold, and J. Evans, “Performance analysis for optical OFDM transmission in short-range IM/DD systems,” J. Lightwave Technol. 30(7), 974–983 (2012). [CrossRef]

6. D. J. F. Barros, S. K. Wilson, and J. M. Kahn, “Comparison of orthogonal frequency-division multiplexing and pulse-amplitude modulation in indoor optical wireless links,” IEEE Trans. Commun. 60(1), 153–163 (2012). [CrossRef]

7. F. Yang, J. Gao, and S. Liu, “Novel visible light communication approach based on hybrid OOK and ACO-OFDM,” IEEE Photonics Technol. Lett. 28(14), 1585–1588 (2016). [CrossRef]

8. F. Zafar, D. Karunatilaka, and R. Parthiban, “Dimming schemes for visible light communication: the state of research,” IEEE Trans. Wireless Commun. 22(2), 29–35 (2015). [CrossRef]

9. X. Ling, J. Wang, X. Liang, Z. Ding, and C. Zhao, “Offset and power optimization for dco-ofdm in visible light communication systems,” IEEE Trans. Signal Process. 64(2), 349–363 (2016). [CrossRef]

10. Y. Sun, F. Yang, and J. Gao, “Comparison of hybrid optical modulation schemes for visible light communication,” IEEE Photon. J. 9(3), 1–13 (2017). [CrossRef]

11. S. D. Dissanayake and J. Armstrong, “Comparison of ACO-OFDM, DCO-OFDM and ADO-OFDM in IM/DD systems,” J. Lightwave Technol. 31(7), 1063–1072 (2013). [CrossRef]

12. B. Ranjha and M. Kavehrad, “Hybrid asymmetrically clipped OFDM-based IM/DD optical wireless system,” J. Opt. Commun. Netw. 6(4), 387–396 (2014). [CrossRef]

13. Q. Wang, C. Qian, and X. Guo, “Layered ACO-OFDM for intensity-modulated direct-detection optical wireless transmission,” Opt. Express 23(9), 12382–12393 (2015). [CrossRef]

14. D. Tsonev, S. Videv, and H. Haas, “Unlocking spectral efficiency in intensity modulation and direct detection systems,” IEEE J. Sel. Areas Commun. 33(9), 1758–1770 (2015). [CrossRef]

15. H. Marshoud, V. M. Kapinas, G. K. Karagiannidis, and S. Muhaidat, “Non-orthogonal multiple access for visible light communications,” IEEE Photonics Technol. Lett. 28(1), 51–54 (2016). [CrossRef]

16. C. Hu, C. Chen, M. Guo, Y. Yang, J. Luo, and L. Chen, “Optical-spatial-summing-based NOMA with fine-grained power allocation for VLC-enabled iot applications,” Opt. Lett. 45(17), 4927–4930 (2020). [CrossRef]

17. Z. Ding, Z. Yang, P. Fan, and H. V. Poor, “On the performance of non-orthogonal multiple access in 5G systems with randomly deployed users,” IEEE Signal Process. Lett. 21(12), 1501–1505 (2014). [CrossRef]

18. Z. Dong, T. Shang, Q. Li, and T. Tang, “Differential evolution-based optimal power allocation scheme for NOMA-VLC systems,” Opt. Express 28(15), 21627–21640 (2020). [CrossRef]

19. X. Guan, Q. Yang, Y. Hong, and C. C.-K. Chan, “Non-orthogonal multiple access with phase pre-distortion in visible light communication,” Opt. Express 24(22), 25816–25823 (2016). [CrossRef]

20. H. Ren, Z. Wang, S. Du, Y. He, J. Chen, S. Han, C. Yu, C. Xu, and J. Yu, “Performance improvement of NOMA visible light communication system by adjusting superposition constellation: a convex optimization approach,” Opt. Express 26(23), 29796–29806 (2018). [CrossRef]

21. H. Li, Z. Huang, Y. Xiao, S. Zhan, and Y. Ji, “Solution for error propagation in a NOMA-based VLC network: symmetric superposition coding,” Opt. Express 25(24), 29856–29863 (2017). [CrossRef]

22. X. Huang, F. Yang, and J. Song, “Novel heterogeneous attocell network based on the enhanced ADO-OFDM for VLC,” IEEE Commun. Lett. 23(1), 40–43 (2019). [CrossRef]

23. S. M. R. Islam, N. Avazov, O. A. Dobre, and K. Kwak, “Power-domain non-orthogonal multiple access (NOMA) in 5G systems: Potentials and challenges,” IEEE Commun. Surveys Tuts. 19(2), 721–742 (2017). [CrossRef]

24. J. Armstrong, “OFDM for optical communications,” J. Lightwave Technol. 27(3), 189–204 (2009). [CrossRef]

25. F. Miramirkhani and M. Uysal, “Channel modeling and characterization for visible light communications,” IEEE Photonics J. 7(6), 1–16 (2015). [CrossRef]

Flexible NOMA-based NOHO-OFDM scheme for visible light communication with iterative interference cancellation

Abstract

1. Introduction

2. NOHO-OFDM with subcarrier allocation

3. Transmitter and receiver structures

3.1 Transmitter structure

3.2 Receiver structure

3.3 Analysis for the achievable data rate

4. Experiments and simulations

5. Conclusions

Funding

Disclosures

References

Cited By

Figures (11)

Equations (13)

Optics Express