Optical Wireless Communication (OWC) is a new trend in communication systems to achieve large bandwidth, high bit rate, high security, fast deployment, and low cost. The basic idea of the OWC is to transmit data on unguided media with light. This system requires multi-carrier modulation such as Orthogonal Frequency Division Multiplexing (OFDM). This paper studies optical OFDM performance based on Intensity Modulation with Direct Detection (IM/DD) system. This system requires a non-negativity constraint. The paper presents a framework for wireless optical OFDM system that comprises (IM/DD) with different forms, Direct Current biased Optical OFDM (DCO-OFDM), Asymmetrically Clipped Optical OFDM (ACO-OFDM), Asymmetrically DC-biased Optical OFDM (ADO-OFDM), and Flip-OFDM. It also considers channel coding as a tool for error control, channel equalization for reducing deterioration due to channel effects, and investigation of the turbulence effects. The evaluation results of the proposed framework reveal enhancement of performance. The performance of the IM/DD-OFDM system is investigated over different channel models such as AWGN, log-normal turbulence channel model, and ceiling bounce channel model. The simulation results show that the BER performance of the IM/DD-OFDM communication system is enhanced while the fading strength is decreased. The results reveal also that Hamming codes, BCH codes, and convolutional codes achieve better BER performance. Also, two algorithms of channel estimation and equalization are considered and compared. These algorithms include the Least Squares (LS) and the Minimum Mean Square Error (MMSE). The simulation results show that the MMSE algorithm gives better BER performance than the LS algorithm.

Optical Wireless Communication (OWC) requires lightwave carriers such as Infra-Red (IR), visible, and Ultra-Violet (UV) light for transmitting data through unguided propagation media. Many ancient cultures use smoke, beacon fires, ship flags, and semaphore telegraph for signaling as examples of historical OWC [

The objective of any modern communication system is to provide high data rates and a wide range of services such as videophones, voice communications, and high-speed Internet access. Orthogonal Frequency Division Multiplexing (OFDM) is an attractive Multi-Carrier Modulation (MCM) technique that efficiently utilizes the available bandwidth. The OFDM systems can transmit high-speed data transmission across a noisy channel and combat multipath propagation effects. So, it is used in many applications such as Digital Audio Broadcasting (DAB) and Terrestrial Digital Video Broadcasting (DVB-T). It is also used in some of the most prominent wireless technologies, such as the IEEE 802.11 Wireless Local Area Networks (WLANs) and Long Term Evolution (LTE) technology [

Recently, OFDM has been applied to optical communication for supporting high data rates. However, the conventional OFDM technique cannot be directly used in optical systems. In OWC systems, the IM/DD technique is a simple, common, and low-cost optical carrier modulation and demodulation technique. In general, it is known that the output of the conventional OFDM modulator is complex and bipolar. In optical systems, only the intensity of the signal is constrained to be real and positive. Therefore, as commonly used in RF communication systems, the conventional OFDM must be modified to be used in OWC systems. Many existing OFDM modulation techniques are suitable for IM/DD OWC systems, such as ACO-OFDM, DCO-OFDM, ADO-OFDM, and Flip-OFDM. These four OFDM-based schemes are discussed in this work.

The rest of the work is organized as follows. Section 2 describes the intensity modulation direct detection (IM/DD) optical wireless communication system architecture. The general IM/DD system is described in Section 3. The channel coding method to improve the BER performance of any communication system is introduced in Section 4. Section 5 illustrates the proposed system models. Simulation results and discussions are given in Section 6. Finally, conclusions are given in Section 7.

In the DCO-OFDM system, data symbols are carried on all sub-carriers [

To prevent any residual DC component or any DC shift in the signal, _{0}, _{N/}_{2} are set to zero and do not carry any information [

In the case of using a large number of sub-carriers _{D}^{2} defined as,

The DC bias level is chosen to be proportional to the root mean square (RMS) of the signal to minimize the amount of required optical power and avoid increasing DC bias [_{DC}_{D}_{DC}

Any remaining negative peaks of _{DC}_{c}_{DC}_{DCO}

The ACO-OFDM technique has been explained in detail in many papers [

In the ACO-OFDM system, the front-end of the ACO-OFDM transmitter is like the DCO-OFDM transmitter, where the output of the IFFT is first converted from P/S, and the CP is appended to it [

The processing in the ACO-OFDM receiver is similar to that of the DCO-OFDM receiver, except that in the ACO-OFDM receiver, only odd sub-carriers that carry data are demodulated, but in DCO-OFDM, all sub-carriers are demodulated.

The ADO-OFDM considers an advanced technique that maintains the advantages and avoids the drawbacks of the two previous optical unipolar OFDM techniques. The ADO-OFDM is a hybrid structure of ACO-OFDM and DCO-OFDM, where a DC bias is added to a part of the signal and the other part of the signal is clipped at zero. Thus, ACO-OFDM symbols modulate odd sub-carriers, while DCO-OFDM symbols modulate even sub-carriers. After that, the negative values produced by ACO-OFDM and DCO-OFDM are separately clipped to zero. Then, the resultant two non-negative signals from ACO-OFDM and DCO-OFDM are added together and transmitted by a LED [

The concept of Flip-OFDM is inverting the polarity of the negative part of the signal. The output of the IFFT operation is a real bipolar signal that can be decomposed into [^{+}^{−}^{+}^{−}_{CP}_{CP}

In the IM/DD system model, the LED emits the transmitted optical power

The ceiling bounce model is the most used model for simulating the impulse response of the indoor OWC channel. It was proposed by Trenkwalder et al. [_{RMS}_{c}

A weak atmospheric turbulence regime is characterized by a single scattering event and can be represented by a single scattering process (Rytov approach). The unitless Rytov variance parameter

There are many techniques, which can be used to improve the Bit Error Rate (BER) performance of any communication system. One of these techniques is channel coding or error detection and correction. In this section, channel coding is used for enhancing the BER performance of DCO-OFDM, ACO-OFDM, ADO-OFDM, and Flip-OFDM in different optical wireless communication channels. Furthermore, hamming code, BCH code, and convolutional codes are proposed in this work to mitigate the noise effect [

Hamming code is a linear block code, where the encoder input is a group of bits with length _{min}

The BCH is the abbreviation for three names of scientists who developed these codes. BCH codes are as consider one of the most efficient codes in linear block coding techniques. They form a class of cyclic error-correcting codes that are constructed using polynomials over a finite field.

There is a binary (^{m } ^{m}_{min}

The BCH code can correct a number of errors equal to or less than

Convolutional codes are different from block codes. Their construction is dependent on using many shift registers, which are used as a memory to store the previous data inputs for calculating the data outputs. The redundant bits are generated in the convolutional coder by using modulo

The DCO-OFDM technique is simple to implement, but the added DC bias makes it inefficient in optical power. Also, this technique suffers from the clipping noise due to the hard clipping of the remaining negative part. The clipping noise may degrade the system performance, especially when low bias levels and large constellation sizes are used.

The block diagram of the proposed DCO-OFDM system with channel coding is shown in _{k}

In ACO-OFDM, no DC bias is required to convert a bipolar signal to a unipolar signal, so that the ACO-OFDM is more efficient than the DCO-OFDM. The signal is clipped at zero levels without adding any clipping noise and without missing any information.

Firstly, the input binary data is encoded by adding redundant data bits using a type of channel coding such as convolutional codes. The encoded data is serial-to-parallel conversion and mapping using (BPSK, QPSK,

To get a real-time valued signal from the IFFT, the input vector to the IFFT,

Like the DCO-OFDM, the IFFT output signal is serially converted, and CP is appended. Then, it is D/A converted and passed through an ideal LPF, resulting in _{ACO}

The generated ACO-OFDM signal is defined as follows:

The proposed ADO-OFDM is the same as the ADO-OFDM except for the decoding process. At the end of the proposed ADO-OFDM receiver, the received ADO-OFDM signal is demodulated and decoded.

The matrix of the MMSE equalizer is given by:
^{H}_{mat}

The performance analysis of the communication system has been estimated by finding the relation between BER and SNR. The BER has been computed by finding the ratio between the number of bit errors (_{ERR}_{TS}

The number of bit errors is obtained by comparing the received bit sequence with the transmitted bit sequence.

AWGN Channel

The BER performance of DCO-OFDM is studied at two different values of DC biasing; 7 and 13 dB. Four cases of QAM constellation mapping are used: 4, 16, 64, and 256 QAM. The simulation parameters are 600 OFDM symbols and 1024 subcarriers in each OFDM symbol and a CP length of 24. The simulation results are shown in

In ^{–4}, convolutional codes with code rates 1/2 and 2/3 enhance the performance by approximately 5 dB. Also, a 7 dB performance enhancement is obtained in the case of the code with a rate of 1/3 at BER = 10^{−4}.

In the case of using low-bias DCO-OFDM, the clipping noise dominates for larger constellation sizes by increasing the BER, which leads to degradation in the system performance. Convolutional codes are used to eliminate clipping noise. Therefore, the BER performance enhancement is achieved. In _{b}_{0} in the case of using convolutional codes. For example, at _{b}_{0} = 25 dB, un-coded DCO-OFDM for 64 QAM gives a BER ≅ 0.13, and convolutional coded DCO-OFDM for 64 QAM gives a BER ≅ 0.0001.

Log-Normal Turbulence Channel

The performance of the OWC system is sensitive to atmospheric turbulence. The DCO-OFDM IM/DD system is affected by weak atmospheric turbulence when used in free space, as shown in

The effect of weak atmospheric turbulence can be mitigated by using error-correcting codes. The Hamming (7, 4) code can be used for improving the performance of the DCO-OFDM system over the log-normal turbulence channel model. The results of these simulation experiments are shown in _{l}^{2 }=^{ }0.01, and 0.001.

AWGN Channel

The BER performance of the ACO-OFDM is studied over the AWGN channel for QAM constellation sizes: 4, 16, 32, 64, 256, and 1024. Convolutional codes with rates 1/2 and 1/3 are used for performance enhancement. The BER performance of the ACO-OFDM is investigated over the AWGN channel with varying QAM constellation sizes, as shown in

The improvement of BER performance of 4 QAM ACO-OFDM over AWGN channel using convolutional codes with rates 1/2 and 1/3 is demonstrated in ^{−3}, it is seen that convolutional codes with rates 1/2 and 1/3 give better performance by approximately 4.5 and 6 dB, respectively.

Log-Normal Turbulence Channel Model

This section studies the BER performance of the ADO-OFDM with BPSK modulation over log-normal turbulence channels. The effect of very weak atmospheric turbulence on the ADO-OFDM system is illustrated in ^{−4}.

_{l}^{2 }=^{ }0.001. Convolutional codes enhance the performance in cases of fading strength σ_{l}^{2} = 0.01, and 0.001. At BER = 10^{−3},

AWGN Channel

This section shows Matlab simulation results for Flip-OFDM. First, the BER performance of Flip-OFDM is studied over the AWGN channel with different modulation types: BPSK, QPSK, and QAM, as shown in ^{−3}, the proposed Flip-OFDM with convolutional code with rate 1/2 provides improvement by approximately 3.5 dB.

The BER performance of un-coded Flip-OFDM and the proposed Flip-OFDM with BCH coding for 4, and 16 QAM over AWGN channel are shown in ^{−3}, the proposed Flip-OFDM with BCH coding enhances 5 dB for 4 and 16 QAM.

Ceiling Bounce Channel Model

^{−3}, the MMSE equalizer outperforms the LS equalizer by nearly 9 dB for 64 QAM Flip-OFDM.

^{−3}, the MMSE equalizer outperforms the LS equalizer by 8 dB for 64 QAM Flip-OFDM.

^{−3}, respectively.

Log-Normal Turbulence Channel Model

_{l}^{2} on BER performance of Flip-OFDM. _{l}^{2 }=^{ }0.1, using convolutional code with a code rate = 1/2, and a constraint length = 3.

This paper studied the performance of OWC IM/DD systems based on four forms of unipolar OFDM with different communication channel models. The objective of the work presented in the thesis is to enhance the performance of these systems using different techniques. This study was divided into three parts. Firstly, the BER performance of the DCO-OFDM, ACO-OFDM, ADO-OFDM, and Flip-OFDM IM/DD system was studied over the AWGN channel. Hamming, BCH, and convolutional codes have been used for performance enhancement. The coded systems were compared with the un-coded systems, and the performance enhancement with coding was elaborated. Secondly, the BER performance of the IM/DD OFDM system was studied over the ceiling bounce channel model. Two channel estimation and equalization algorithms have been considered and compared: the LS algorithm and the MMSE algorithm. The simulation results show that the MMSE algorithm gives better BER performance than the LS algorithm. Then, convolutional code was utilized for BER performance enhancement. Finally, the effect of weak atmospheric turbulence was studied on IM/DD OFDM system by studying the BER performance of the system over the log-normal turbulence channel model. The obtained results show that the weak atmospheric turbulence gradually decreases the BER performance of the IM/DD OFDM system with an increase in fading strength. Also, Hamming, BCH, and convolutional codes were utilized to mitigate the effect of very weak atmospheric turbulence. For future work, the suggested signal processing tools could be employed in advanced 5G communication networks with different types of modulation techniques.

The authors would like to thank the support of the Deanship of Scientific Research at Princess Nourah bint Abdulrahman University.