Optical orthogonal frequency division multiplexing index modulation (OOFDM-IM) is a novel multicarrier technique that achieves higher transmission rates by additionally adding mode order indexing via loading two different modes of constellation symbols on the active subcarriers and retaining the silent subcarriers. At the same time, the presence of silent subcarriers improves the bit error rate (BER) performance of the system. However, the presence of silent subcarriers also brings a waste of spectrum resources. Therefore, in this paper, a zero-padded dual-mode optical orthogonal frequency division multiplexing index modulation (ZDM-OOFDM-IM) is proposed to improve the transmission rate. Meanwhile, in order to solve the problem of excessive system detection complexity, a three-level stepwise detection algorithm based on K-mean clustering algorithm (KMC++) is proposed by combining the greedy (GD) algorithm, log-likelihood ratio (LLR) algorithm, and KMC++ algorithm.

In the ZDM-OOFDM-IM system, the active subcarrier was selected by the subcarrier index, and the constellation symbols of two different modes were mapped by the symbol information. Then, the modulation symbols were loaded on the active subcarriers according to the order index to complete signal modulation. After orthogonal frequency division multiplexing (OFDM) data block merging, inverse fast Fourier transform, non-zero clipping, and other processing, the light source sent it out. The optical signal transmitted through the atmospheric turbulence channel was received by the detector and could be restored to the original signal after processing by fast Fourier transform, subcarrier recovery, and maximum likelihood (ML) detection. According to the subcarrier index, constellation pattern order, and different characteristics of constellation symbols in the ZDM-OOFDM-IM system, GD, LLR, and KMC++ algorithms were used to detect the proposed detection algorithm, so as to reduce the complexity of the receiver signal detection. Finally, the feasibility of the system and the proposed algorithm were verified by the Monte Carlo method and experimental equipment.

In this paper, the ZDM-OOFDM-IM system and its low-complexity novel detection algorithm are proposed, and its feasibility is verified through simulation and experimental devices. In addition, the influence of the key parameters of the system on the BER is analyzed. The results show that the ZDM-OOFDM-IM system can effectively improve the transmission rate, which gradually approaches the dual-mode OOFDM-IM with the increase in the subcarrier block length and the number of active subcarriers (Fig. 2). Moreover, at all modulation orders of 4, the signal-to-noise ratio (SNR) of the proposed system is improved by about 4.02 dB compared to DM-OOFDM-IM at BER of 3.8×10^-3 (Fig. 4). In addition, when the subcarrier block length is fixed, increasing the number of active subcarriers can significantly increase the transmission rate but inevitably results in BER loss. When the number of active subcarriers is constant, an increase in subcarrier block length leads to an improvement in BER performance. For example, compared to the (4,2,1,2,4) system, the transmission rate of one frame of the (4,3,1,2,4) and (4,3,2,2,4) systems is improved by 64 bit/s and 128 bit/s, respectively, while their SNRs are lost by 2.17 dB and 2.26 dB at BER of 3.8×10^-3, respectively. Compared with the (4,2,1,2,4), (4,3,1,2,4), and (4,3,2,2,4) systems at BER of 3.8×10^-3, the SNRs of the (8,2,1,2,4), (8,3,1,2,4), and (8,3,2,2,4) systems are improved by 4.53 dB, 3.72 dB, and 2.50 dB, respectively (Fig. 5). The proposed algorithm achieves a BER that approximates the ML detection and eliminates the “plateau effect” of the conventional KMC algorithm (Fig. 6). Under the proposed algorithm, it is concluded that the increase in the modulation order leads to a significant increase in the transmission rate although it brings a smaller BER loss. For example, when ML detection is employed at BER of 3.8×10^-3, the SNRs of the (4,2,1,2,2) and (4,2,1,4,4) systems lose 3.41 dB and 5.56 dB, respectively, compared to the (4,2,1,2,2) system, whereas the transmission rates of their one-frame signals are enhanced by 64 bit/s and 128 bit/s, respectively (Fig. 7). Moreover, in the experimental setup, the system BER and the performance of the proposed algorithm also achieve results consistent with the simulation (Fig. 9 and Fig. 10). Finally, the computational complexity of the proposed algorithm is given and compared with several classical decoding algorithms to demonstrate its computational complexity advantage (Fig. 11).

In order to solve the problem of unsatisfactory transmission rate and BER performance in the traditional wireless OOFDM-IM system, a ZDM-OOFDM-IM system is designed in this paper, which effectively enhances the transmission rate. Compared with the OOFDM-IM system, its transmission rate is increased by 96 bit/s for one frame of information when the number of subcarriers is 16, and the number of active subcarriers is 8. Meanwhile, in order to solve the problems of excessive ML decoding complexity and poor BER performance of other traditional detection algorithms, a three-stage stepwise detection algorithm is proposed by combining the GD, LLR, and KMC++ algorithms, which obtains a complexity close to that of linear decoding algorithms under the premise of guaranteeing the BER performance. Finally, the system successfully realizes wireless optical communication transmission with a BER of lower than 3.8×10-3 through the constructed indoor transmission experimental device, which verifies the feasibility of the system and the proposed algorithm.