International Journal of Performability Engineering, 2018, 14(12): 3228-3236 doi: 10.23940/ijpe.18.12.p32.32283236

A Rotated Constellation Aided OFDM System for Wireless Communication

Qianqian Luo,, Huiheng Liu, and Lixin Song

School of Physics and Electronic Engineering, Hubei University of Arts and Science, Xiangyang,441053, China

*Corresponding Author(s): * E-mail address: 11082439@qq.com

Accepted:  Published:   

Abstract

The most common way to improve spectral efficiency of the orthogonal frequency-division multiplexing (OFDM) communication system is to increase the modulation orders. However, studies have shown that this method will increase the bit error rate and symbol error rate simultaneously. In order to improve the spectral efficiency of OFDM system without reducing its reliability, this paper proposes a rotated four leaves clover (R-FLC) constellation. In this scheme, each symbol can carry additional bits through alternately employing two rotated constellations with different phases. A general rotated constellation aided OFDM system for wireless communication is also proposed. The bit error rate and effective spectral efficiency performances of this system under additional white Gaussian noise and multipath Rayleigh fading channel are evaluated in detail. Compared with the traditional Quadrature Amplitude Modulation (QAM) aided OFDM system, simulation results show that the proposed system is capable of improving the spectral efficiency without sacrificing signal-to-noise ratio under additional white Gaussian noise channel and multipath Rayleigh fading channel. The proposed system can effectively overcome the Inter Symbol Interference (ISI). The system architecture has a certain reference and practical value for the development of OFDM modulation technology.

Keywords: rotated four leaves clover; OFDM; signal-to-noise ratio; effective spectral efficiency

PDF (276KB) Metadata Related articles Export EndNote| Ris| Bibtex

Cite this article

Qianqian Luo, Huiheng Liu, Lixin Song. A Rotated Constellation Aided OFDM System for Wireless Communication. International Journal of Performability Engineering, 2018, 14(12): 3228-3236 doi:10.23940/ijpe.18.12.p32.32283236

1. Introduction

With the development of economy and technology, the requirement of information transmission rate and service quality becomes higher and higher. However, the scarcity of spectrum resource limits the further development of wireless communication. The openness of the wireless channel and the time-varying characteristics of channel parameters also bring serious damage to the transmission of wireless signals [1]. With the characteristics of high spectral efficiency (SE) and anti-multipath interference capability, etc., OFDM technology [2] can satisfy the demand of wireless communication for high rate, wide broadband and mobility. It has become the core modulation transmission technology of the next generation broadband wireless communication system.

OFDM [3-6] is a particular multicarrier transmission technology. The main idea of it is to transform high rate information into parallel low rate symbols and then transmit them parallel in a number of orthogonal sub-carriers, thereby reducing the effect of frequency selective fading (FSF). In OFDM, the entire channel is divided into many narrow sub-channels which can be utilized in parallel transmission; the carriers of the sub-channels are orthogonal to each other resulting in overlap of the subcarriers in frequency spectrum, thereby reducing the effect of mutual interference between the subcarriers and greatly improving the SE simultaneously. How to improve the SE of the OFDM communication system has been a main research focus. The most common way is to increase the modulation orders; however, doing so will increase the bit error rate (BER) and symbol error rate (SER) simultaneously. SE and BER seem to be a pair of mutually exclusive performance indexes. At present, several modulation methods have been proposed to get better data transfer rate for improving SE. A rotated quadrature phase shift keying (R-QPSK) technique was proposed in [7]; this scheme can transmit 3 bits per symbol period, effectively achieving higher throughput albeit at the cost of increased complexity. An improved scheme was proposed in [8] and [9] where one additional bit can be transmitted by adopting two R-QPSK constellations with different phases. This technique can get the gain in throughput without sacrificing BER. Based on [9], Ref. [10] proposed a quad state-paired QPSK (QS-PQPSK) scheme. Simulation results showed that its performance is better than 8PSK.

Inspired by [8] and [9], this paper proposes a rotated four leaves clover (R-FLC) constellation whose Euclidean distance is larger than 8PSK; therefore, its BER performance is better than 8PSK. Adopting the new constellation, the system can transmit additional bits by employing rotation mechanism; therefore, the SE can be improved. In addition, we apply the rotated constellations mechanism into the traditional OFDM system, putting forward a universal rotated constellation aided OFDM system. Simulation results show that the comprehensive performance of this system is much better than the traditional quadrature amplitude modulation (QAM) based OFDM system.

The remaining part of this paper is organized as follows: In section 2, after a brief review on how to transmit additional bits by R-QPSK constellation, an R-FLC constellation is proposed. Then, a rotated constellation aided OFDM system model is introduced in Section 3, where the SE of this new system is discussed. In Section 4, extensive simulation results about BER and ESE performances are given and analyzed. Finally, Section 5 concludes the whole work.

2. R-FLC Constellation

Ref. [8] and [9] proposed an R-QPSK constellation scheme capable of transmitting additional information. As shown in Figure 1, R-QPSK scheme utilizes two QPSK constellations: Q0 and Q1. The phase difference between Q0 and Q1 is 45 degrees and the Minimum Euclidean distance (MED) of those two constellations is same.

Figure 1

Figure 1.   R-QPSK constellation proposed by [9]


The basic principle of R-QPSK is as follows. In the transmitter, 2Nbits information are modulated by different constellations Qk, where $k\in \left\{ 0,1 \right\}$, constellation Q0 represent sone bit information “0”, while constellation Q1 represents the other bit information “1”.Thus, the transmitter can send (2N+1) bits information duringN symbol periods.

In Figure 1, the constellation points $S_{{{Q}_{k}}}^{n}$ of Q0 and Q1 do not overlap with each other in the Euclid space, where $n\in \left\{ 1,\text{ }2,\text{ }3 \right\}$, thereby making the receiver detect and determine the received signal belonging to Q0orQ1. The extraction of the additional bits carried by the group of information can be determined by comparing the accumulated distance between constellation Qk and the received N symbols.

Inspired by the scheme of R-QPSK constellation, a rotated 8 points constellation is proposed. As shown in Figure 2, the shape of this constellation is similar to the four leaves clover (FLC); therefore, we name it the R-FLCconstellation.

Figure 2

Figure 2.   A rotated four leaves clover (R-FLC) constellation


From Figure 2, it can be seen that any two adjacent points in the inner circle and a point in the big circle constitute a regular triangle. This arrangement can acquire a lager MED than 8PSK in the condition of normalized average power, resulting inan improved BER under the high SNR condition.

The MEDs of the constellations are provided in Table 1. As shown, the MED of R-FLC constellation is improved by 20% more than 8PSK.

Table 1.   MEDs of different signal constellations t

ConstellationMEDs
R-QPSK Q0 or Q11.414
R-FLC Q0 or Q10.9193
8PSK0.7653

New window| CSV


Similar to R-QPSK, R-FLC employs a pair of FLC constellations Q0 and Q1 that have a phase difference of 45 degreesbetween each other; the MEDs of the two constellations are same. The transmitter sends 3Nbits information that were modulated by the same type of R-FLC constellationQk, where $k\in \left\{ 0,\text{ }1 \right\}$. Therefore, the transmitter can send (3N+1) bits only occupying N symbol periods. The extraction of the additional bits is also decided by comparing the accumulated distance between constellation Qk and the received symbols.

3. Rotated Constellation Aided OFDM System Model

3.1. System Model

In this section, the rotated constellation is applied into the traditional OFDM transmission system. The transmitter of this new OFDM system is shown in Figure 3. Set the subcarrier number of this system as N. After serials/parallel (S/P) conversion, the data sequence is transformed into N roads original data that will be sent to the rotated constellation mapper module. The input data sequence X forming an OFDM symbol can be represented as

$\begin{matrix} & X=\left( {{x}_{b,0}},{{x}_{b,1}},\cdots ,{{x}_{b,N-1}} \right) \\ & =\left( {{b}_{0}}\cdots {{b}_{{{N}_{b}}-1}},{{b}_{{{N}_{b}}}}\cdots {{b}_{2{{N}_{b}}-1}},\cdots ,{{b}_{\left( N-1 \right){{N}_{b}}}}\cdots {{b}_{N{{N}_{b}}-1}} \right) \end{matrix}$

Where ${{x}_{b,m}}=({{b}_{m{{N}_{b}}}},\cdots ,{{b}_{(m+1){{N}_{b}}-1}})$, $m\in \{0,1,\cdots ,N-1\}$, Nb is the modulation order of the constellation.

Figure 3

Figure 3.   Rotated constellation aided OFDM transmitter


As shown in Figure 3, in this system, each OFDM symbol can carry additional bits information. That is because N sub-channels are equally divided into N' groups. The sub-channel number of each group is Nd=N/N', where ${{N}_{d}}\in {{Z}^{+}}$. The sub-channels of the same group employ the same mapping constellation Qk so that an OFDM symbol can carry N' bits without occupying time slot. The additional bits stream carried by an OFDM symbol can be represented as

$A=\left( {{a}_{0}}{{a}_{1}}\cdots {{a}_{N'-1}} \right)$

Where N' represents the length of the additional bits.

After S/P conversion of A, the single bit of A can determine which constellation the Nd sub-channels of the ith group utilize. After $X=\left( {{x}_{b,0}},{{x}_{b,1}},\cdots ,{{x}_{b,}}_{N-1} \right)$ is mapped by rotated constellation, N complex signal $S=\left( {{S}_{0}},{{S}_{1}},\cdots ,{{S}_{N-1}} \right)$ can be obtained. Then, S is converted into time-domain signal $S=\left( {{S}_{0}},{{S}_{1}},\cdots ,{{S}_{N-1}} \right)$ by N points inverse discrete Fourier transform (IDFT). Finally, a parallel/serials (P/S) conversion is employed to form an OFDM symbol. The OFDM symbol will be transmitted into the wireless channel after adding some cyclic redundancy prefix (CP) which can effectively reduce the effect of inter-symbol interference (ISI).

In the receiver, by discrete Fourier transformation (DFT), an OFDM symbol is generated into N complex signal $Y=\left( {{y}_{0}},{{y}_{1}},\cdots ,{{y}_{N-1}} \right)$.Y is equally divided into N' groups and re-written as a matrix YN'×Nd,which can be expressed as

${{Y}_{N'\times {{N}_{d}}}}=\left[ \begin{matrix} {{y}_{0,0}} & {{y}_{0,1}} & ... & {{y}_{0,{{N}_{d}}-1}} \\ {{y}_{1,0}} & {{y}_{1,1}} & ... & {{y}_{1,{{N}_{d}}-1}} \\ ... & ... & ... & ... \\ {{y}_{N'-1,0}} & {{y}_{N'-1,1}} & ... & {{y}_{N'-1,{{N}_{d}}-1}} \\ \end{matrix} \right]$

Wherethe matrix element yi,j represents the received jth symbol of the ith data group, $i\in [0,N'-1]$ and $j\in [0,{{N}_{d}}-1]$.

According to the system model, each row and/or group of the matrix can carry 1-bit additional information. If we can distinguish which constellation used by the group, the corresponding additional information of each group can be decoded correctly. Determining constellation type also utilizes comparing the accumulated distance of the two types of constellations and the Nd symbols of the same group. The accumulated distance between the ith data group and constellation Qk can be expressed as in

${{d}_{k,i}}=\sum\limits_{j=0}^{{{N}_{d}}-1}{\min \left\| {{y}_{i,j}}-{{Q}_{k}} \right\|},\ \ \ k=0,1$

The estimation and decision of the additional information ${{\hat{a}}_{i}}$ of the ith data group is given by

${{\hat{a}}_{i}}=\left\{ \begin{matrix} 0, & \text{if }{{d}_{0,i}}\le {{d}_{1,i}} \\ 1, & \text{if}\text{ }{{d}_{0,i}}>{{d}_{1,i}} \\ \end{matrix} \right.$

The demodulation of the symbol yi,j is as follows. The corresponding constellation type is firstly determined by the row number i of the yi,j. Then, we adopt maximum-likelihood criterion to demodulate the original data $\gamma $.$\gamma$ can be expressed as

$\gamma =\underset{n=1,2,\cdots ,{{2}^{{{N}_{b}}}}}{\mathop{\arg \min }}\,\left\| {{y}_{i,j}}-s_{{{Q}_{k}}}^{n} \right\|$

Where $S_{{{Q}_{k}}}^{n}$ represents the nth constellation point of the corresponding constellation type Qk of the ith row symbol.

3.2. Analysis of Spectrum Efficiency (SE)

SE is an important technical indicator of OFDM system. Assuming the sampling time is TS seconds, we have the duration time and the bandwidth of an OFDM symbol as NTS seconds and 1/TS Hz respectively; the transmission rate RS of the OFDM symbol of this system is 1/NTSsymbols per second. In our system, we know that the data rate R of an OFDM symbol is $N{{N}_{b}}+N'$ bits per symbol; thus, the Spectrum efficiency $\eta $ of this system can be expressed as

$\eta =\frac{R{{R}_{S}}}{W}=\frac{N{{N}_{b}}+N'}{N}={{N}_{b}}+\frac{N'}{N}$

From (7), we can see that the larger the ratio of N' to N, the larger SE will be.

4. Simulation and Analysis

In this section, performances of this proposed rotated constellation aided OFDM system under AWGN are evaluated and simulated, and the BER performances of R-QPSK and R-FLC under different numbers of additional bits N' are compared. The number of subcarriers N is 120.

Figure 4 is BER performances of the R-QPSK aided OFDM system under AWGN where N' is the number of additional bits carried by an OFDM symbol. From Figure 4, it can be seen that when $N'\le 12$, the BER performance of R-QPSK is close to the conventional QPSK and the SE has a slightly increase than QPSK. When N'=20 and N'=30, the BER performance of R-QPSK is slightly worse than QPSK in small Eb/No and is slightly better than QPSK in large Eb/No. When N'=40 and N'=60, the R-QPSK lose 0.7dB and 1.7dB Eb/No gain than QPSK respectively at the BER of ${{10}^{-3}}$. In conclusion, R-QPSK can obtain at most 12.5% SE gain than QPSK without sacrificing Eb/No.

Figure 4

Figure 4.   BER performances of the R-QPSK aided OFDM system under AWGN


Figure 5 is the BER performances of the R-FLC aided OFDM system under AWGN where N' is also the number of additional bits carried by an OFDM symbol. From Figure 5, it can be seen that when $N'\le 40$, the BER performances of R-FLC are better than 8PSK. The new system can obtain at least 1dB Eb/No gain at the BER of ${{10}^{-3}}$ and the SEs of R-FLC also spontaneously improved. When N'=60, the BER performance of R-FLC is slightly worse than 8PSK in small Eb/No; however, it is close to the 8PSK in large Eb/No. Simulation results indicate that spectrum efficiency can be improved by 16.7% than 8PSK without sacrificing Eb/No gain when using the R-FLC scheme.

Figure 5

Figure 5.   BER Performances of the R-FLC aided OFDM system under AWGN


To directly and comprehensively evaluate the rotated aided OFDM system, a comprehensive metric known as effective spectral efficiency (ESE)is proposed; it can be expressed as

${{\eta }_{E}}=\left( 1-{{P}_{FER}} \right)\frac{{{R}_{F}}}{{{T}_{F}}W}$ (bit/s/Hz)

Where RF is the information bits per frame, TF is the time duration per frame, RFER is the frame error rate.

Assuming a frame includes NF OFDM symbols, according to (7), we can get ${{R}_{F}}=(N{{N}_{b}}+N'){{N}_{F}}$ bits per frame and ${{T}_{F}}={{N}_{F}}N{{T}_{s}}$; thus, (8) can be represented as

${{\eta }_{E}}=\left( 1-{{P}_{FER}} \right)\frac{N{{N}_{b}}+N'}{N}=\left( 1-{{P}_{FER}} \right)({{N}_{b}}+\frac{N'}{N})$ (bit/s/Hz)

As seen from (9), we simulate ESEs of R-QPSK and R-FLC in our simulation. The frame size NF is set as 1000, i.e. a frame contains 1000 OFDM symbols.

Figure 6 is the ESE curves of R-QPSK aided OFDM system under AWGN according to (9). It can be seen that when $N'\le 30$, the ascending curve of R-QPSK is close to the QPSK; this phenomenon demonstrates that the SE of R-QPSK architecture has an improvement of QPSK without the loss of SNR gain. It is further proved that the comprehensive performance of R-QPSK is better than QPSK. When N'=40, the R-QPSK loses about 1dB SNR gain than QPSK but acquires 16.7% SE gain. When N'=60, the R-QPSK loses about 2dB SNR gain than QPSK but acquires 25% SE gain.

Figure 6

Figure 6.   ESEs of the R-QPSK aided OFDM system under AWGN


Figure 7 is the ESE curves of R-FLC aided OFDM system under AWGN according to (9). From Figure 7, it can be seen that when $N'\le 40$, the ascending curve of R-FLC is at the left of the QPSK. It demonstrates that R-FLC not only obtains 1dB SNR gain than 8PSK, but also acquires the SE gain. When N'=60, the ascending curve of R-FLC coincides with the QPSK; this phenomenon shows that the SE has a 16.7% improvement without the loss of SNR gain.

Figure 7

Figure 7.   ESEs of the R-FLC aided OFDM system under AWGN


In the following, the BER performance of the rotated constellations aided OFDM system under multipath Rayleigh channel is investigated and evaluated. The multipath Rayleigh channel may arouse Inter Symbol Interference (ISI); therefore, we need add cyclic prefix (CP) to prevent ISI. Assuming the length of CP is${{N}_{c}}$, the duration of an OFDM symbol is$({{N}_{c}}+N){{T}_{s}}$. According to (7), the SE of the new OFDM system adding CP can be represented as

${{\eta }_{c}}=\frac{R{{R}_{S}}}{W}=\frac{N{{N}_{b}}+N'}{N+{{N}_{c}}}$ (bit/s/Hz)

In the simulations, the number of subcarriers used in the OFDM system is 120, while the number of the CP is set to 16.

Figure 8 is the BER performances of the R-QPSK aided OFDM system with different number of additional bits N' under multi-path Rayleigh fading channel. When N'=12, the R-QPSK obtains about 1dB Eb/No gain than QPSK at the BER of 10-3; in additional, the R-QPSK acquires 5% SE gain. When N'=30, the R-QPSK is close to the conventional QPSK; however, the SE of R-QPSK has a 12.5% improvement than QPSK. In conclusion, R-QPSK can obtain at most 12.5% SE gain than QPSK without sacrificing the Eb/No gain under multi-path Rayleigh fading channel.

Figure 8

Figure 8.   BER Performances of the R-QPSK aided OFDM system under multi-path Rayleigh fading channel


Figure 9 is BER performances of the R-FLC aided OFDM system under MULTI-PATH RAYLEIGH FADING CHANNEL. From the simulation results, it can be seen that when N'=30, R-FLC obtains about 0.2dB Eb/No gain than 8PSK at the BER of ${{10}^{-3}}$;in addition, the R-FLC also acquires8.33% SE gain. When N'=60, the R-FLC is close to the conventional 8PSK; however, the SE of R-FLC has 16.7% improvement than 8PSK. In conclusion, R-FLC can obtain at most 16.7% of SE gain than 8PSK without the loss of Eb/No gain under multi-path Rayleigh fading channel.

Figure 9

Figure 9.   BER Performances of the R-FLC aided OFDM system under multi-path Rayleigh fading channel


5. Conclusions

A rotated four leaves clover constellation and a general rotated constellation aided OFDM system are provided in this paper. We compare the comprehensive performance of this new system with the conventional QAM aided OFDM system at the same modulation level. Extensive simulation results demonstrate that the SE has an improvement without sacrificing the SNR gain. The system architecture has a certain reference and practical value for the development of OFDM modulation technology.

Acknowledgements

This work was supported by the Scientific Research Project of Education Department of Hubei Province, China (No. B2017160).

Reference

L. Dai, J. Wang, Z. Wang, P. Tsiaflakis, M. Moonen , “

Spectrum-and Energy-Efficient OFDM based on Simultaneous Multi-Channel Reconstruction

,” IEEE Transactions on Signal Processing, Vol. 61, No. 23, pp. 6047-6059, 2013

[Cited within: 1]

A. U. Ahmed and J. R. Zeidler, “

Novel Low-Complexity Receivers for Constant Envelope OFDM

,” IEEE Transactions on Signal Process, Vol. 63, No. 17, pp. 4572-4582, 2015

[Cited within: 1]

S. W. Hong, E. S. Kang, and D. S. Han , “

Additional Data Transmission with Rotated QPSK Constellation

,” Electronics Letters, Vol. 51, No. 5, pp. 394-395, 2015

[Cited within: 1]

C. Wang, G. Cui, W. Wang , “

Dual-Stream Transceiver Structure with Single Antenna for Phase-Modulated OFDM

,” IEEE Communications Letters, Vol. 20, No. 9, pp. 1756-1759, 2016

T. Wang, J. G. Proakis, E. Masry, J. R. Zeidler , “

Performance Degradation of OFDM Systems Due to Doppler Spreading

,” IEEE Transactions on Wireless Communications, Vol. 5, No. 6, pp. 1422-1432, 2006

Z. H. E. Tan and A. K. Krishna , “

A Quad State-Paired QPSK Modulation for Higher Data Rate Communication

,” in Proceedings of IEEE 83rd Vehicular Technology Conference (VTC Spring), Nanjing, 2016

[Cited within: 1]

Z. Zhang, F. Tian, L. Ge, F. Zeng, G. Xuan , “

M-Ary Spread Spectrum OFDM Structure based on Cascaded Cyclic Shift Complementary Pairs

,” in Proceedings of International Conference on Signal Processing IEEE, pp. 1584-1589, Kerala, 2015

K. Davaslioglu and E. Ayanoglu, “

Quantifying Potential Energy Efficiency Gain in Green Cellular Wireless Networks

,” IEEE Communications Surveys & Tutorials, Vol. 16, No. 4, pp. 2065-2091, 2014

[Cited within: 2]

J. Liu, J. Kim, S. Kwatra, and G. Stevens ,“

Rotative Quadrature Phase shift Keying

,” Electronics Letters, Vol.28, No.12, pp. 1095, 1992

[Cited within: 3]

E. S. Kang, S. W. Hong, D. S. Han , “

Improved Speed near Field Communication with Rotated QPSK Constellation and Hidden Data Transmission

,” in IEEE international Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), pp. 1-2, Seoul, 2012

/